Envelope-to: ccamp-data@psg.com Delivery-date: Sun, 30 Sep 2001 23:52:04 -0700 Message-Id: <4.3.2.7.2.20011001023523.023ef400@mo-ex1> Date: Mon, 01 Oct 2001 02:47:44 -0400 To: ccamp@ops.ietf.org From: Lou Berger Subject: update GMPLS signaling documents Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed FYI - Updated versions of the signaling and RSVP-TE documents have been submitted. The CR-LDP version is forthcoming. The updates address comments received during the August meeting and comment period. Here's a summary of changes: draft-ietf-mpls-generalized-signaling-06.txt Changes from previous version: o Removed redundant encoding type o Revised Admin Status Usage (per comments and also removed potential race condition) o Clarified text related to interface bundling (To be consistent with updated bundling draft.) o Minor editorial changes and clarifications o Revised G-PID (per comments on list) draft-ietf-mpls-generalized-rsvp-te-05.txt Changes from previous version: o Revised Admin Status Usage (per comments and also removed potential race condition) o Clarified text related to interface bundling (To be consistent with updated bundling draft.) o Added IF_ID ERROR_SPEC Object (Per comments, supports Control Channel Separation by allowing identification of errored interface.) o Added U bit to Label RRO subobject (Per comment, to match defined ERO.) o Clarified usage of Restart_Cap o Replaced use of Suggested_Label during recovery with new Recovery_Label (Per last WG session) o Added IANA Considerations o Minor editorial changes and clarifications For those who don't want to wait, the drafts are available from: http://www.labn.net/docs/draft-ietf-mpls-generalized-signaling-06.txt http://www.labn.net/docs/draft-ietf-mpls-generalized-rsvp-te-05.txt Lou Envelope-to: ccamp-data@psg.com Delivery-date: Sun, 30 Sep 2001 01:24:37 -0700 From: "manoj juneja" To: ccamp@ops.ietf.org Bcc: Date: Sun, 30 Sep 2001 01:20:28 -0700 Mime-Version: 1.0 Content-Type: text/plain; format=flowed Message-ID: Hi All, If minimum reservable bandwidth of an STM-16 link in interface switching capability TLV is VC-3, does this mean it can't switch VC -2/VC-11 etc ? In other words, if a request comes for establishing an LSP with signal type as VC-11, it can't establish this LSP ? STM-16 has 16 branches, if each of these branches support different min. reservable bandwidth then does this mean it will be having 16 interface switching capability TLVs ? Is this a deviation from SDH/SONET where u can't assume that if minimum reservable bandwidth is say VC-2 and max. reservable bandwidth is say STM-1, it doesn't mean that all the signals viz. VC-4, VC-12, VC-3 etc. (i.e. that fall under STM-1 and VC-2) are supported ? Please correct me if I am wrong. Regards, manoj. _________________________________________________________________ Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 28 Sep 2001 18:45:45 -0700 From: "manoj juneja" To: ccamp@ops.ietf.org Bcc: Subject: Bi-directional LSPs [Contention Resolution] Date: Fri, 28 Sep 2001 18:41:36 -0700 Mime-Version: 1.0 Content-Type: text/plain; format=flowed Message-ID: Hi All, I have a basic doubt regarding contention resolution. The draft explains the example with upstream label and suggested label. Suggested label is optional. If a node is not able to allocate the suggested label then it may override the suggestged label. The actual contention arises when it is the last label. Am I right ? Another case, the contention will also occur when say request for two bi-directional LSPs are initiated in opposite directions and both the requests include same upstream label. In addition to this, the label set contains one label, same as upstream label (i.e. sender wants to have same label in both the directions). Please clarify. Regards, manoj. _________________________________________________________________ Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 27 Sep 2001 09:06:33 -0700 Message-ID: <87009604743AD411B1F600508BA0F959632464@XOVER.dedham.mindspeed.com> From: "Atkinson, Stephen" To: ccamp@ops.ietf.org Subject: Questions on draft-ietf-mpls-bundle-mib-00 Date: Thu, 27 Sep 2001 12:04:27 -0400 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" I have a few questions on the bundling mib. 1. Is the data-bearing channel table defined in draft-ietf-mpls-bundle-mib-00 supposed to represent a Component as defined in draft-ietf-mpls-bundle-00? 2. If the answer to 1 above is "yes", how can you specify a Component's IP address or Component ID? 3. Perhaps specifying a Component as defined in draft-ietf-mpls-bundle-00 requires both a data-bearing channel and a teLink? If this is the case, what is the significance of multiple data-bearing channels under a single teLink? Thanks, steve atkinson Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 25 Sep 2001 09:29:15 -0700 Message-ID: From: Jonathan Lang To: 'George Young' Cc: ccamp@ops.ietf.org Subject: RE: Question: purpose of LMP ChannelActive message (was relations hip of RSVP-TE and LMP) Date: Tue, 25 Sep 2001 09:07:48 -0700 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" George, Sorry for the delayed response. Please see inline. Thanks, Jonathan > -----Original Message----- > From: George Young [mailto:george.young@edgeflow.com] > Sent: Thursday, September 20, 2001 5:52 AM > To: Jonathan Lang > Cc: ccamp@ops.ietf.org > Subject: RE: Question: purpose of LMP ChannelActive message (was > relationship of RSVP-TE and LMP) > > > Hello Jonathan, > > Thanks again for the clarification. There's something that still > troubles me. > > I can see that if there's signalling involved (e.g. RSVP), then the > signalling indicates whether monitoring is necessary along the path > (e.g. with a couple of bits in the Admin Status object) and the LMP > ChannelActive message is superfluous. > > If there's no signalling involved, then the LMP ChannelActive message > could be used to indicate monitoring should take place (e.g. for span > protection). The trouble is that when monitoring finds that a channel > has failed and sends a ChannelFail message, the transmit end knows to > switch over, but without the ChannelFailNack message, the receive end of > the link doesn't know to switch over. > > Am I missing something here? One thing we're doing for the next rev of LMP is combining the ChannelFail, ChannelActive, ChannelDeactive messages into a single ChannelStatus message (as discussed in LMP-WDM). With this in place, you could send a ChannelStatus (fail) message to indicate the failure. > > Regards, > George. > edgeflow Inc. > 329 March Rd., Kanata, ON, Canada, K2K 2E1 > phone: +1 613-270-9279 Ext 287 > fax: +1 613-270-9268 > > > >-----Original Message----- > >From: Jonathan Lang [mailto:jplang@calient.net] > >Sent: Wednesday, September 19, 2001 1:38 PM > >To: George Young; ccamp@ops.ietf.org > >Subject: RE: Question: purpose of LMP ChannelActive message (was > >relationship of RSVP-TE and LMP) > > > > > >George, > > see inline. > > > >Thanks, > >Jonathan > > > >> -----Original Message----- > >> From: George Young [mailto:george.young@edgeflow.com] > >> Sent: Wednesday, September 19, 2001 8:42 AM > >> To: Jonathan Lang; ccamp@ops.ietf.org > >> Subject: RE: Question: purpose of LMP ChannelActive message (was > >> relationship of RSVP-TE and LMP) > >> > >> > >> Thanks for the clarification, Jonathan. > >> > >> To see if I understand, the active call and the backup call are > >> separately established by something like RSVP-TE. These > two calls are > >> established by PATH going from head to tail, RESV going > from tail to > >> head. Then for the active call ONLY, the head sends a ChannelActive > >> which propagates to the tail, and then the tail sends a > ChannelActive > >> which propagates to the head. > >> > >> When failure is noted by monitoring the active call, the > head and the > >> tail of the call are notified, and each switches over to the backup > >> call. The head sends a ChannelActive message downstream > and the tail > >> sends a ChannelActive message upstream, and they're each > >propagated to > >> the other end. > >> > >> Right? > >not quite right. I think you're confusing the ChannelActive > >message with > >the Administrative Status object in GMPLS. The ChannelActive > >message is > >used between a pair of nodes to indicate that the data channel > >is now active > >and should be monitored. This is useful for span level protection or > >between devices where GMPLS signaling (RSVP) isn't used. > > > >> > >> Regards, > >> George. > >> edgeflow Inc. > >> 329 March Rd., Kanata, ON, Canada, K2K 2E1 > >> phone: +1 613-270-9279 Ext 287 > >> fax: +1 613-270-9268 > >> > >> > >> >-----Original Message----- > >> >From: Jonathan Lang [mailto:jplang@calient.net] > >> >Sent: Tuesday, September 18, 2001 12:24 PM > >> >To: George Young; ccamp@ops.ietf.org > >> >Subject: RE: Question: purpose of LMP ChannelActive message (was > >> >relationship of RSVP-TE and LMP) > >> > > >> > > >> >George, > >> > Sorry, I missed this one. See inline. > >> > > >> >-Jonathan > >> > > >> >> -----Original Message----- > >> >> From: George Young [mailto:george.young@edgeflow.com] > >> >> Sent: Tuesday, September 18, 2001 6:05 AM > >> >> To: ccamp@ops.ietf.org > >> >> Subject: Question: purpose of LMP ChannelActive message (was > >> >> relationship of RSVP-TE and LMP) > >> >> > >> >> > >> >> Didn't see any answers to this question, so I'll try to > >> rephrase and > >> >> make it more focused. > >> >> > >> >> What is the purpose of the LMP ChannelActive message? > >> >The ChannelActive message is used to indicate that the channel > >> >status is > >> >active and should be monitored. > >> > > >> >> > >> >> I'm assuming a call is set up through an optical cross > >> connect (OXC), > >> >> either in response to control plane signalling (e.g. > RSVP-TE) or > >> >> management plane input. > >> >> The OXC has knowledge of which optical links are involved > >> >> with the call for failure monitoring purposes. > >> >> > >> >> What additional information would the receipt of a > >> >> ChannelActive message convey? > >> >This is useful when paths are pre-established (e.g., for path > >> >protection). > >> >Although signaling is used to setup the resource reservation, > >> >on switchover > >> >you need a mechanism to trigger monitoring of the data channel. > >> > > >> >> > >> >> Regards, > >> >> George. > >> >> edgeflow Inc. > >> >> 329 March Rd., Kanata, ON, Canada, K2K 2E1 > >> >> phone: +1 613-270-9279 Ext 287 > >> >> fax: +1 613-270-9268 > >> >> > >> >> > >> >> >-----Original Message----- > >> >> >From: George Young > >> >> >Sent: Friday, September 07, 2001 3:47 PM > >> >> >To: ccamp@ops.ietf.org > >> >> >Subject: relationship of RSVP-TE and LMP > >> >> > > >> >> > > >> >> >I have two questions about the relationship of RSVP-TE > >> >> >(draft-ietf-mpls-generalized-rsvp-te-03.txt) and LMP > >> >> >(draft-ietf-ccamp-lmp-00.txt). > >> >> > > >> >> >First of all, my understanding of the LMP ChannelActive > >> >> message is that > >> >> >it is used to indicate to the downstream (relative to data flow > >> >> >direction) node that monitoring should be used. > >> >> > > >> >> >The scenario I have in mind is an RSVP call setup > consisting of a > >> >> >bidirectional optical path. Presumably a switching node has > >> >> to reserve > >> >> >data channels and associate them with the optical path when an > >> >> >RSVP PATH > >> >> >message arrives. > >> >> > > >> >> >1) My first question is what is the timing of LMP > >> >> >ChannelActive messages > >> >> >relative to RSVP messages. > >> >> > > >> >> >Do ChannelActive messages flow downstream (relative to call > >> >> >establishment direction) more-or-less in phase with PATH > >> >> messages, and > >> >> >upstream (relative to call establishment direction) > >> more-or-less in > >> >> >phase with RESV messages? > >> >> > > >> >> >2) My second question is what addtional information > >information do > >> >> >ChannelActive messages provide? > >> >> > > >> >> >If a link is to be monitored when associated with an > >> active call, it > >> >> >appears the node already has this information. > >> >> > > >> >> >Regards, > >> >> >George. > >> >> >edgeflow Inc. > >> >> >329 March Rd., Kanata, ON, Canada, K2K 2E1 > >> >> >phone: +1 613-270-9279 Ext 287 > >> >> >fax: +1 613-270-9268 > >> >> > > >> >> > > >> >> > > >> >> > >> > > >> > > > Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 25 Sep 2001 09:21:08 -0700 Message-ID: From: Jonathan Lang To: "'Bhargava, Nidhi'" , 'Fugui Wang' , John Drake , "'kireeti@juniper.net'" , "'yakov@juniper.net'" , "'lberger@movaz.com'" , "'dsaha@tellium.com'" , "'dbasak@accelight.com'" , "'hsandick@nortelnetworks.com'" , "'braja@tellium.com'" Cc: "'ccamp@ops.ietf.org'" Subject: RE: Questions regarding draft-ietf-ccamp-lmp-00.txt (LMP) Date: Tue, 25 Sep 2001 09:06:02 -0700 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Nidhi, Sorry for the delay in responding. Please see inline. Thanks, Jonathan > -----Original Message----- > From: Bhargava, Nidhi [mailto:nidhi.bhargava@netplane.com] > Sent: Thursday, September 20, 2001 6:45 AM > To: Jonathan Lang; 'Fugui Wang'; John Drake; 'kireeti@juniper.net'; > 'yakov@juniper.net'; 'lberger@movaz.com'; 'dsaha@tellium.com'; > 'dbasak@accelight.com'; 'hsandick@nortelnetworks.com'; > 'braja@tellium.com' > Cc: 'ccamp@ops.ietf.org' > Subject: RE: Questions regarding draft-ietf-ccamp-lmp-00.txt (LMP) > > > Jonathan, > With reference to answer to question (2), do we wait for the other side > so set the control channel down bit in the hello messages before going to > DOWN state or do we immediately go to DOWN state without caring about the > neighbour. I will add the following to the next rev of the draft: "The node may stop sending Hello messages after HelloDeadInterval seconds have passed, or if it receives an LMP message over the same control channel with the ControlChannelDown flag set." > > Also I have a question regarding the reconfig event. When the reconfig > event is generated, do we always send the config message regardless of > neighbours node id. Or we will only send config is our node if is greater > than neighbours. Do we need to stop the hello protocol before doing the > renegotiation? Yes, you can send the config message (regardless of neighbor's node ID) if you want to reconfig. You can continue sending the Hello protocol. > > Lastly when the control channel reboots, we are expected to come back to > the UP state and resume Hello protocol. Shouldn't we be doing a config > procedure again. Does the node reboot bit in the common header represents the > node reboot or the control channel reboot? You should do a config procedure again. The node reboot bit has been renamed to LMP Restart to indicate that the LMP module has restarted. > > Thanks > Nidhi > ------------------------------- > Nidhi Bhargava > NetPlane Systems Inc. > A Mindspeed Technologies Company > Tel: + 1.781.329.3200 x5353 > Web: http://www.netplane.com > > > > > > > > -----Original Message----- > > From: Jonathan Lang [mailto:jplang@calient.net] > > Sent: Wednesday, September 19, 2001 6:40 PM > > To: 'Fugui Wang'; John Drake; 'kireeti@juniper.net'; > > 'yakov@juniper.net'; 'lberger@movaz.com'; 'dsaha@tellium.com'; > > 'dbasak@accelight.com'; 'hsandick@nortelnetworks.com'; > > 'braja@tellium.com' > > Cc: 'ccamp@ops.ietf.org' > > Subject: RE: Questions regarding draft-ietf-ccamp-lmp-00.txt (LMP) > > > > > > Fugui, > > See inline. > > > > Thanks, > > Jonathan > > > > > -----Original Message----- > > > From: Fugui Wang [mailto:fwang@axiowave.com] > > > Sent: Wednesday, September 19, 2001 8:42 AM > > > To: Jonathan Lang; Krishna Palla; John Drake; > 'kireeti@juniper.net'; > > > 'yakov@juniper.net'; 'lberger@movaz.com'; 'dsaha@tellium.com'; > > > 'dbasak@accelight.com'; 'hsandick@nortelnetworks.com'; > > > 'azinin@cisco.com'; 'braja@tellium.com' > > > Cc: 'ccamp@ops.ietf.org' > > > Subject: Questions regarding draft-ietf-ccamp-lmp-00.txt (LMP) > > > > > > > > > Hi, > > > > > > During the implementation of the LMP protocol, I have got several > > > questions regarding the LMP draft. Could you please clarify any of > > > them? > > > > > > Thanks, > > > Fugui > > > > > > 1. Control Channel FSM: > > > > > > (1) There are two events which could bring up the control channel, > > evBringUp_1a, > > > evBringUp_1b. Does that mean that in the MIB we need to > > have something > > like > > > activeUp and pasvUp in the AdminStatus of control channel table? > > I don't think this is necessary. > > > > > > > > (2) In this version, the GoingDn state is removed. When a > > control channel > > > gets evAdminDown event, it will go to Down state immediately. > > > Should it send a Hello or other messages with > > ControlChannelDown bit set > > to the > > > neighbor before it goes down? Since the Hello message is sent > > periodically. If we > > > don't send an additional Hello with ControlChannelDown > bit set, most > > probably, the > > > control channel would never have a chance to send such a > > message. Then the > > remote > > > node will get Hello time out instead of get the event evNbrGoesDn. > > We will add text to the evAdminDown event indicating that > > Hello messages > > sent before going down should have the ControlChannelDown bit set. > > > > > > > > (3) When receiving a ConfigNack, if the Hello Config TLV is > > nonegotiable > > or > > > it is negotiable but we could not find the values acceptable for > > > both side, should we still keep sending Config message? > > According to the > > > Control Channel FSM, when getting the evConfErr, the > control channel > > should keep in > > > ConfSnd state. > > If the configuration parameters are negotiable, and the > > proposed values are > > acceptable, a new Config message should be transmitted with > > these values. > > If the Nack'd values are non-negotiable, or if the proposed > values are > > unacceptable, you can go into the down state (event 2) or > > keep trying with > > your values (hoping that configuration is changed at the far end). > > > > > > > > 2. TE link FSM: > > In the next version of LMP we are simplifying the TE link > > FSM. The TE Link > > can really only be in the UP, DOWN, and DEGRADED states. > > Link Verification > > can occur in either the UP or DOWN states, and could be > > represented using > > sub-states. > > > > > > > > (1) After verification is done (evVerDone), the FSM goes > > from VrfProc to > > > Summary state. However, if all of the Data Bearing links > > failed in the > > > verification, should the TE link still enter the Summary > > state and send a > > > LinkSummary message with no TLV? Or should it go from > > VrfProc to Down > > state > > > directly without sending such a meaningless LinkSummary > > message? (This > > could > > > happen if the Data Bearing links and TE links have not been > > configured in > > > the remote side yet and the local side has already > > configured and start to > > > verify.) > > > > > > (2) In the Summary state, the node keep sending LinkSummary > > message to > > remote > > > until it receives LinkSummaryAck or LinkSummaryNack. Should > > we have an > > other > > > events such as evSummaryTimeout so that we won't keep > > trying infinetely if > > > we couldn't get any response from the remote? > > > > > > (3) There is no event in the FSM indicating what will > > happen if the node > > > receives BeginVerify, LinkSummary. My understanding is that > > the receiver > > > side is stateless. Is that correct? > > > > > > (4) How to handle BeginVerifyNack? There is no event in the FSM > > > corresponding to this. > > > > > > 3. BeginVerify Message: > > > > > > The BeginVerify message brings a field VerifyInterval. But my > > understanding > > > is that the receiver side should care more about the Verify Dead > > > Interval than the Verify Interval since it needs to use the > > Verify Dead > > > Interval to determine when to send TestStatusFailure back. > > (Note that the > > > reciever side may not be able to get the Verify Dead > > Interval from the > > local TE link > > > configuration If the sender side don't know the remote TE > > link ID and want > > to learn > > > it through verification.) > > The VerifyDeadInterval is determined at the receive side and > > must take into > > account the VerifyInterval at which the Test messages are sent. > > > Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 24 Sep 2001 08:54:36 -0700 Message-Id: <4.3.2.7.2.20010924114502.03f5af08@bucket.cisco.com> Date: Mon, 24 Sep 2001 11:48:47 -0400 To: Irfan Lateef From: Thomas Nadeau Subject: Re: OSPF-GMPLS-EXTN MIB Cc: ccamp@ops.ietf.org, mpls@UU.NET Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed At 12:55 PM 9/21/2001 -0400, Irfan Lateef wrote: >Tom, > >I should have read this draft more closely, but thanks for pointing >to the tables which links the pieces to together. I does the job well. > >However I have a few comments. > >1. As mentioned in section.3 of the draft "bundled link is another kind of >TE link" > and the MIB does indeed address other kinds of links( like GMPLS links > not necessaryly using a bundling concept). The title of the draft >deserves a > reconsideration which encompasses the a generalized notion. Let me cut >the story > short, consider renaming it "TE GMPLS and Bundled Link Mgmt MIB" or >something. The name was chosen to reflect the draft that it models objects for. The working group accepted the document named as it is. >2. The MIB ideally models the bundling concept and the GMPLS attributes >seems to > be an after thought, I understand the necessity to re-use the existing >stuff. From a > GMPLS point of view, the attributes appear to be scattered around four >different > tables. I am not sure how you may want to do it. But there is need to >group them to together in someway. Please point out where you think there are deficiencies and how you would like us to resolve them. >My suggestion is, there can be a conformance >group "GMPLS link group" similar to linkBundlingGroup which can help is >identifying how the MIB can be used to model GMPLS links which are not >ecessaryly bundled. I don't understand how a conformance group can help in this regard. Conformance groups are used to point out which objects must be supported and which are optional (and under what circumstances). >3. Some description about the relationship with OSPF MIB can help. Will do. >4. Although the draft does point to [GMPLS-OSPF] in sec.3 but a explicit >sentence in > section.2 like "The same objects can also be used to model GMPLS links > and their attributes as defined [GMPLS-OSPF]", would great help. Okay. --Tom >All these comments are editorial in nature but may help others in future who > >are trying to grapple with GMPLS . > >Regards, >Irfan Lateef > >Thomas Nadeau wrote: > > > Take a look at the Link Bundling MIB, > > specifically the teLinkOspfTeTable. The MIB can be > > found here: > > > > http://search.ietf.org/internet-drafts/draft-ietf-mpls-bundle-mib-00.txt > > > > --Tom > > > > >This is with reference to the OSPF extension to GMPLS > > > draft-ietf-ccamp-ospf-gmpls-extensions-00.txt. > > > > > > I know OSPF has corresponding MIB, OSPF-MIB (RFC 1850). > > > Is there a corresponding MIB for the ospf-gmpls-extn > > > which extends OSPF-MIB? > > > If the answer is no, How can the NMS find the additional > > > attributes of the GMPLS links. > > > Did I miss something! > > > Appreciate your response. > > > > > > Irfan Lateef > > > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 24 Sep 2001 03:07:07 -0700 Message-ID: <3BAF0449.6000806@npd.hcltech.com> Date: Mon, 24 Sep 2001 15:30:41 +0530 From: S Ramesh Organization: HCL Technologies Ltd. User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.2) Gecko/20010726 Netscape6/6.1 MIME-Version: 1.0 To: ccamp@ops.ietf.org Subject: Explicit Label Control -RSVP-TE ? Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Hi, In ERO the U- bit definition is not clear in draft-ietf-mpls-generalized-rsvp-te-04.txt. Can anybody explain the U-bit set and unset (0 & 1). Specifically my understanding is that if U-bit is 0 then the label is considered as incoming label and the outgoing label must be allocated on the outgoing interface and copied into upstream label object and sent with the outgoing PATH message. I cannot understand when U-bit is one. Kindly explain in depth. TIA, Ramesh Envelope-to: ccamp-data@psg.com Delivery-date: Sun, 23 Sep 2001 07:33:49 -0700 Message-Id: <200109231428.f8NESV018205@merlot.juniper.net> To: "Naidu, Venkata" cc: "'Yakov Rekhter'" , "'ccamp@ops.ietf.org'" , "'mpls@uu.net'" Subject: Re: GMPS Unnumbered MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-ID: <50886.1001255311.1@juniper.net> Date: Sun, 23 Sep 2001 07:28:31 -0700 From: Yakov Rekhter Venkata, > Yakov: > > -> > * MPLS-TE extensions define Interface IP Address TLVs > -> > (Type 3 and 4) as length of 4N. Where N is the number of > -> > local IP addresses for that Interface. > -> > > -> > But in GMPLS, why can't we configure multiple "unique" > -> > Interface IDs for Unnumbered interfaces? Please allow > -> > flexibility with out restricting one ID for Unnumbered Link. > -> > -> What kind of problem(s) would that solve ? > > > If for example, a physical p2p link is sub divided into > multiple "logical" links (such as ATM VP/VC) each has > identifiers but represents same other characteristics > (like bw etc) then there is no point in generating multiple > link TLVs for each of those "logical" interfaces. Considering > OSPF is implemented as per draft-ietf-ospf-ppp-flood-01.txt. > (hope you won't suggest link bundling here) what would be the problem(s) with using link bundling in this case ? > Please see more comments on GMPLS Routing below: > > * There is no mention about Interface MTU in > draft-ietf-ccamp-gmpls-routing-00.txt, but I can see the > use in draft-ietf-ccamp-ospf-gmpls-extensions-00.txt > (as Sub TLV type 10). > > More over, "logical TE link" (with no OSPF adjacency) need not > signify "Interface MTU" but "path MTU" suites well(?) I actually wonder whether it would make more sense to carry MTU not as a Sub-TLV, but in the Switching Capability specific information (for PSC links only). (For one thing, I don't think MTU has any meaning on non-PSC links). > * I suggest/request to include "delay" also in GMPLS enhancements. > None of the TE extensions (MPLS, DiffServ, GMPLS) considered > "link delay" as one of path computation constraint. (could be > NP-complete but there are means to compute path and very helpful > in Traffic Management/QoS cases for delay sensitive applications) Did you have a chance to look at the proposal suggested in draft-lefaucheur-te-metric-igp-00.txt ? > * Is that Sub-TLV types 3|4 and 11|12 are mutually exclusive? > If not what would be the behavior if OSPF receives all of > above TLVs. (is there any possibility here that we can save > some types/overhead with regards to scalability) I guess in principle the same link could be both numbered and unnumbered (why would one do this in practice is a completely different question). So, at least in principle one could argue that 3|4 and 11|12 may not be mutually exclusive. > > * Minor: there is no mention about which Sub-TLVs are optional > and which are mandatory. And any restriction about processing > the (un)recognized TLVs? This draft just follows the approach of the "basic" MPLS TE routing extensions. Yakov. Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 21 Sep 2001 09:54:05 -0700 Message-ID: <3BAB70F7.87358166@vill.com> Date: Fri, 21 Sep 2001 12:55:19 -0400 From: Irfan Lateef Organization: Village Networks Inc MIME-Version: 1.0 To: Thomas Nadeau CC: ccamp@ops.ietf.org, mpls@uu.net Subject: Re: OSPF-GMPLS-EXTN MIB Content-Type: multipart/mixed; boundary="------------F826DB3D2DA0C2D123E0E536" This is a multi-part message in MIME format. --------------F826DB3D2DA0C2D123E0E536 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Tom, I should have read this draft more closely, but thanks for pointing to the tables which links the pieces to together. I does the job well. However I have a few comments. 1. As mentioned in section.3 of the draft "bundled link is another kind of TE link" and the MIB does indeed address other kinds of links( like GMPLS links not necessaryly using a bundling concept). The title of the draft deserves a reconsideration which encompasses the a generalized notion. Let me cut the story short, consider renaming it "TE GMPLS and Bundled Link Mgmt MIB" or something. 2. The MIB ideally models the bundling concept and the GMPLS attributes seems to be an after thought, I understand the necessity to re-use the existing stuff. From a GMPLS point of view, the attributes appear to be scattered around four different tables. I am not sure how you may want to do it. But there is need to group them to together in someway. My suggestion is, there can be a conformance group "GMPLS link group" similar to linkBundlingGroup which can help is identifying how the MIB can be used to model GMPLS links which are not necessaryly bundled. 3. Some description about the relationship with OSPF MIB can help. 4. Although the draft does point to [GMPLS-OSPF] in sec.3 but a explicit sentence in section.2 like "The same objects can also be used to model GMPLS links and their attributes as defined [GMPLS-OSPF]", would great help. All these comments are editorial in nature but may help others in future who are trying to grapple with GMPLS . Regards, Irfan Lateef Thomas Nadeau wrote: > Take a look at the Link Bundling MIB, > specifically the teLinkOspfTeTable. The MIB can be > found here: > > http://search.ietf.org/internet-drafts/draft-ietf-mpls-bundle-mib-00.txt > > --Tom > > >This is with reference to the OSPF extension to GMPLS > > draft-ietf-ccamp-ospf-gmpls-extensions-00.txt. > > > > I know OSPF has corresponding MIB, OSPF-MIB (RFC 1850). > > Is there a corresponding MIB for the ospf-gmpls-extn > > which extends OSPF-MIB? > > If the answer is no, How can the NMS find the additional > > attributes of the GMPLS links. > > Did I miss something! > > Appreciate your response. > > > > Irfan Lateef > > > > --------------F826DB3D2DA0C2D123E0E536 Content-Type: text/x-vcard; charset=us-ascii; name="ilateef.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Irfan Lateef Content-Disposition: attachment; filename="ilateef.vcf" begin:vcard n:Lateef;Irfan tel;fax:732-460-5998 tel;work:732-460-7760 x-mozilla-html:FALSE url:www.villagenetworks.com org:Village Networks Inc;Systems Engineering adr:;;246, Industrial Way;Eatontown;New Jersey;07724;USA version:2.1 email;internet:ilateef@vill.com title:Senior Systems Engineer, fn:Irfan end:vcard --------------F826DB3D2DA0C2D123E0E536-- Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 21 Sep 2001 09:14:13 -0700 Message-ID: <3BAB66A6.8EBDCD9F@nayna.com> Date: Fri, 21 Sep 2001 09:11:18 -0700 From: Sudheer Dharanikota MIME-Version: 1.0 To: Irfan Lateef CC: ccamp@ops.ietf.org Subject: Re: OSPF-GMPLS-EXTN MIB Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Please refer to "draft-ietf-mpls-bundle-mib-00.txt". - sudheer Irfan Lateef wrote: > Hi, > This is with reference to the OSPF extension to GMPLS > draft-ietf-ccamp-ospf-gmpls-extensions-00.txt. > > I know OSPF has corresponding MIB, OSPF-MIB (RFC 1850). > Is there a corresponding MIB for the ospf-gmpls-extn > which extends OSPF-MIB? > If the answer is no, How can the NMS find the additional > attributes of the GMPLS links. > Did I miss something! > Appreciate your response. > > Irfan Lateef Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 21 Sep 2001 07:12:26 -0700 Message-ID: <39469E08BD83D411A3D900204840EC5554C87A@vie-msgusr-01.dc.fore.com> From: "Naidu, Venkata" To: "'Yakov Rekhter'" Cc: "'ccamp@ops.ietf.org'" , "'mpls@uu.net'" Subject: RE: GMPS Unnumbered Date: Fri, 21 Sep 2001 09:46:55 -0400 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Yakov: -> > * MPLS-TE extensions define Interface IP Address TLVs -> > (Type 3 and 4) as length of 4N. Where N is the number of -> > local IP addresses for that Interface. -> > -> > But in GMPLS, why can't we configure multiple "unique" -> > Interface IDs for Unnumbered interfaces? Please allow -> > flexibility with out restricting one ID for Unnumbered Link. -> -> What kind of problem(s) would that solve ? If for example, a physical p2p link is sub divided into multiple "logical" links (such as ATM VP/VC) each has identifiers but represents same other characteristics (like bw etc) then there is no point in generating multiple link TLVs for each of those "logical" interfaces. Considering OSPF is implemented as per draft-ietf-ospf-ppp-flood-01.txt. (hope you won't suggest link bundling here) Please see more comments on GMPLS Routing below: * There is no mention about Interface MTU in draft-ietf-ccamp-gmpls-routing-00.txt, but I can see the use in draft-ietf-ccamp-ospf-gmpls-extensions-00.txt (as Sub TLV type 10). More over, "logical TE link" (with no OSPF adjacency) need not signify "Interface MTU" but "path MTU" suites well(?) * I suggest/request to include "delay" also in GMPLS enhancements. None of the TE extensions (MPLS, DiffServ, GMPLS) considered "link delay" as one of path computation constraint. (could be NP-complete but there are means to compute path and very helpful in Traffic Management/QoS cases for delay sensitive applications) * Is that Sub-TLV types 3|4 and 11|12 are mutually exclusive? If not what would be the behavior if OSPF receives all of above TLVs. (is there any possibility here that we can save some types/overhead with regards to scalability) * Minor: there is no mention about which Sub-TLVs are optional and which are mandatory. And any restriction about processing the (un)recognized TLVs? I have few more comments with regards to Link Protection and SRLG. Thanks! --Venkata Naidu Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 21 Sep 2001 07:07:33 -0700 Message-Id: <4.3.2.7.2.20010921100457.01d07e50@bucket.cisco.com> Date: Fri, 21 Sep 2001 10:07:12 -0400 To: Irfan Lateef From: Thomas Nadeau Subject: Re: OSPF-GMPLS-EXTN MIB Cc: ccamp@ops.ietf.org Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Take a look at the Link Bundling MIB, specifically the teLinkOspfTeTable. The MIB can be found here: http://search.ietf.org/internet-drafts/draft-ietf-mpls-bundle-mib-00.txt --Tom >This is with reference to the OSPF extension to GMPLS > draft-ietf-ccamp-ospf-gmpls-extensions-00.txt. > > I know OSPF has corresponding MIB, OSPF-MIB (RFC 1850). > Is there a corresponding MIB for the ospf-gmpls-extn > which extends OSPF-MIB? > If the answer is no, How can the NMS find the additional > attributes of the GMPLS links. > Did I miss something! > Appreciate your response. > > Irfan Lateef > > Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 21 Sep 2001 06:57:37 -0700 Message-ID: <3BAB471D.447B6536@vill.com> Date: Fri, 21 Sep 2001 09:56:45 -0400 From: Irfan Lateef Organization: Village Networks Inc MIME-Version: 1.0 To: ccamp@ops.ietf.org Subject: OSPF-GMPLS-EXTN MIB Content-Type: multipart/mixed; boundary="------------649B04FCE8EFC7D545737C64" This is a multi-part message in MIME format. --------------649B04FCE8EFC7D545737C64 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi, This is with reference to the OSPF extension to GMPLS draft-ietf-ccamp-ospf-gmpls-extensions-00.txt. I know OSPF has corresponding MIB, OSPF-MIB (RFC 1850). Is there a corresponding MIB for the ospf-gmpls-extn which extends OSPF-MIB? If the answer is no, How can the NMS find the additional attributes of the GMPLS links. Did I miss something! Appreciate your response. Irfan Lateef --------------649B04FCE8EFC7D545737C64 Content-Type: text/x-vcard; charset=us-ascii; name="ilateef.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Irfan Lateef Content-Disposition: attachment; filename="ilateef.vcf" begin:vcard n:Lateef;Irfan tel;fax:732-460-5998 tel;work:732-460-7760 x-mozilla-html:FALSE url:www.villagenetworks.com org:Village Networks Inc;Systems Engineering adr:;;246, Industrial Way;Eatontown;New Jersey;07724;USA version:2.1 email;internet:ilateef@vill.com title:Senior Systems Engineer, fn:Irfan end:vcard --------------649B04FCE8EFC7D545737C64-- Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 20 Sep 2001 11:04:26 -0700 Message-Id: <200109201802.f8KI2d084606@merlot.juniper.net> To: "Naidu, Venkata" cc: "'ccamp@ops.ietf.org'" , "'mpls@uu.net'" Subject: Re: GMPS Unnumbered MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-ID: <79416.1001008958.1@juniper.net> Date: Thu, 20 Sep 2001 11:02:39 -0700 From: Yakov Rekhter Venkata, > All GMPLS-Unnumbered: > > I have few comments on draft-ietf-ccamp-gmpls-routing-00.txt > (applicable to draft-ietf-ccamp-ospf-gmpls-extensions-00.txt) > > * MPLS-TE extensions define Interface IP Address TLVs > (Type 3 and 4) as length of 4N. Where N is the number of > local IP addresses for that Interface. > > But in GMPLS, why can't we configure multiple "unique" > Interface IDs for Unnumbered interfaces? Please allow > flexibility with out restricting one ID for Unnumbered Link. What kind of problem(s) would that solve ? > There is no scarcity for local significant integers > (unlike IP address) ;-) > > * It would be helpful, if you give more information about "why" > this outgoing unnumbered link ID is chosen as "unique" in the > scope of LSR. > > This is because, IP (Classical) has Unnumbered links defined > in RFC 1812 Section 2.2.7 - which suggests to use "Router ID" > (MAY be one of the Routers IP addresses) as Unnumbered Link ID. > So in IP context this is not a unique ID. > > My question is, can we use the same IP configured unnumbered > link address for GMPLS purpose? unnumbered link ID has to be unique in the scope of an LSR because the LSR needs to identify a particular link (for the purpose of resource and label allocation), and a pair of LSRs may have multiple links between then. Yakov. Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 20 Sep 2001 10:02:41 -0700 Message-ID: <87009604743AD411B1F600508BA0F9598AD906@XOVER.dedham.mindspeed.com> From: "Kullberg, Alan" To: "'ccamp@ops.ietf.org'" Subject: GMPLS RSVP Nodal Fault recovery Date: Thu, 20 Sep 2001 12:57:31 -0400 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" I think I've found a problem with the nodal fault recovery mechanism in draft-ietf-mpls-generalized-rsvp-te-04.txt, section 9.5. If the node that fails is the ingress node (i.e., an LER), then the LSP won't recover since there is no upstream neighbor from which this restarting LER will receive a PATH message. Therefore, the LER won't send a PATH message to its downstream neighbor, and the LSP recovery fails. In other words, recovery only seems to work if an LSR or an Egress LER experiences nodal failure. Perhaps this could be fixed by having neighbors of the restarting node send a RESV for any Resv state held? Alan Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 20 Sep 2001 06:47:08 -0700 Message-ID: <87009604743AD411B1F600508BA0F959F3DEA7@XOVER.dedham.mindspeed.com> From: "Bhargava, Nidhi" To: 'Jonathan Lang' , 'Fugui Wang' , John Drake , "'kireeti@juniper.net'" , "'yakov@juniper.net'" , "'lberger@movaz.com'" , "'dsaha@tellium.com'" , "'dbasak@accelight.com'" , "'hsandick@nortelnetworks.com'" , "'braja@tellium.com'" Cc: "'ccamp@ops.ietf.org'" Subject: RE: Questions regarding draft-ietf-ccamp-lmp-00.txt (LMP) Date: Thu, 20 Sep 2001 09:44:47 -0400 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Jonathan, With reference to answer to question (2), do we wait for the other side so set the control channel down bit in the hello messages before going to DOWN state or do we immediately go to DOWN state without caring about the neighbour. Also I have a question regarding the reconfig event. When the reconfig event is generated, do we always send the config message regardless of neighbours node id. Or we will only send config is our node if is greater than neighbours. Do we need to stop the hello protocol before doing the renegotiation? Lastly when the control channel reboots, we are expected to come back to the UP state and resume Hello protocol. Shouldn't we be doing a config procedure again. Does the node reboot bit in the common header represents the node reboot or the control channel reboot? Thanks Nidhi ------------------------------- Nidhi Bhargava NetPlane Systems Inc. A Mindspeed Technologies Company Tel: + 1.781.329.3200 x5353 Web: http://www.netplane.com > -----Original Message----- > From: Jonathan Lang [mailto:jplang@calient.net] > Sent: Wednesday, September 19, 2001 6:40 PM > To: 'Fugui Wang'; John Drake; 'kireeti@juniper.net'; > 'yakov@juniper.net'; 'lberger@movaz.com'; 'dsaha@tellium.com'; > 'dbasak@accelight.com'; 'hsandick@nortelnetworks.com'; > 'braja@tellium.com' > Cc: 'ccamp@ops.ietf.org' > Subject: RE: Questions regarding draft-ietf-ccamp-lmp-00.txt (LMP) > > > Fugui, > See inline. > > Thanks, > Jonathan > > > -----Original Message----- > > From: Fugui Wang [mailto:fwang@axiowave.com] > > Sent: Wednesday, September 19, 2001 8:42 AM > > To: Jonathan Lang; Krishna Palla; John Drake; 'kireeti@juniper.net'; > > 'yakov@juniper.net'; 'lberger@movaz.com'; 'dsaha@tellium.com'; > > 'dbasak@accelight.com'; 'hsandick@nortelnetworks.com'; > > 'azinin@cisco.com'; 'braja@tellium.com' > > Cc: 'ccamp@ops.ietf.org' > > Subject: Questions regarding draft-ietf-ccamp-lmp-00.txt (LMP) > > > > > > Hi, > > > > During the implementation of the LMP protocol, I have got several > > questions regarding the LMP draft. Could you please clarify any of > > them? > > > > Thanks, > > Fugui > > > > 1. Control Channel FSM: > > > > (1) There are two events which could bring up the control channel, > evBringUp_1a, > > evBringUp_1b. Does that mean that in the MIB we need to > have something > like > > activeUp and pasvUp in the AdminStatus of control channel table? > I don't think this is necessary. > > > > > (2) In this version, the GoingDn state is removed. When a > control channel > > gets evAdminDown event, it will go to Down state immediately. > > Should it send a Hello or other messages with > ControlChannelDown bit set > to the > > neighbor before it goes down? Since the Hello message is sent > periodically. If we > > don't send an additional Hello with ControlChannelDown bit set, most > probably, the > > control channel would never have a chance to send such a > message. Then the > remote > > node will get Hello time out instead of get the event evNbrGoesDn. > We will add text to the evAdminDown event indicating that > Hello messages > sent before going down should have the ControlChannelDown bit set. > > > > > (3) When receiving a ConfigNack, if the Hello Config TLV is > nonegotiable > or > > it is negotiable but we could not find the values acceptable for > > both side, should we still keep sending Config message? > According to the > > Control Channel FSM, when getting the evConfErr, the control channel > should keep in > > ConfSnd state. > If the configuration parameters are negotiable, and the > proposed values are > acceptable, a new Config message should be transmitted with > these values. > If the Nack'd values are non-negotiable, or if the proposed values are > unacceptable, you can go into the down state (event 2) or > keep trying with > your values (hoping that configuration is changed at the far end). > > > > > 2. TE link FSM: > In the next version of LMP we are simplifying the TE link > FSM. The TE Link > can really only be in the UP, DOWN, and DEGRADED states. > Link Verification > can occur in either the UP or DOWN states, and could be > represented using > sub-states. > > > > > (1) After verification is done (evVerDone), the FSM goes > from VrfProc to > > Summary state. However, if all of the Data Bearing links > failed in the > > verification, should the TE link still enter the Summary > state and send a > > LinkSummary message with no TLV? Or should it go from > VrfProc to Down > state > > directly without sending such a meaningless LinkSummary > message? (This > could > > happen if the Data Bearing links and TE links have not been > configured in > > the remote side yet and the local side has already > configured and start to > > verify.) > > > > (2) In the Summary state, the node keep sending LinkSummary > message to > remote > > until it receives LinkSummaryAck or LinkSummaryNack. Should > we have an > other > > events such as evSummaryTimeout so that we won't keep > trying infinetely if > > we couldn't get any response from the remote? > > > > (3) There is no event in the FSM indicating what will > happen if the node > > receives BeginVerify, LinkSummary. My understanding is that > the receiver > > side is stateless. Is that correct? > > > > (4) How to handle BeginVerifyNack? There is no event in the FSM > > corresponding to this. > > > > 3. BeginVerify Message: > > > > The BeginVerify message brings a field VerifyInterval. But my > understanding > > is that the receiver side should care more about the Verify Dead > > Interval than the Verify Interval since it needs to use the > Verify Dead > > Interval to determine when to send TestStatusFailure back. > (Note that the > > reciever side may not be able to get the Verify Dead > Interval from the > local TE link > > configuration If the sender side don't know the remote TE > link ID and want > to learn > > it through verification.) > The VerifyDeadInterval is determined at the receive side and > must take into > account the VerifyInterval at which the Test messages are sent. > Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 20 Sep 2001 05:58:36 -0700 content-class: urn:content-classes:message MIME-Version: 1.0 Subject: RE: Question: purpose of LMP ChannelActive message (was relationship of RSVP-TE and LMP) Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Date: Thu, 20 Sep 2001 08:51:53 -0400 Message-ID: Thread-Topic: Question: purpose of LMP ChannelActive message (was relationship of RSVP-TE and LMP) Thread-Index: AcFBMliGeIEu91ObTu+dkPipORbwpgAnkA9A From: "George Young" To: "Jonathan Lang" Cc: Hello Jonathan, Thanks again for the clarification. There's something that still troubles me. I can see that if there's signalling involved (e.g. RSVP), then the signalling indicates whether monitoring is necessary along the path (e.g. with a couple of bits in the Admin Status object) and the LMP ChannelActive message is superfluous. If there's no signalling involved, then the LMP ChannelActive message could be used to indicate monitoring should take place (e.g. for span protection). The trouble is that when monitoring finds that a channel has failed and sends a ChannelFail message, the transmit end knows to switch over, but without the ChannelFailNack message, the receive end of the link doesn't know to switch over. Am I missing something here? Regards, George. edgeflow Inc. 329 March Rd., Kanata, ON, Canada, K2K 2E1 phone: +1 613-270-9279 Ext 287 fax: +1 613-270-9268 >-----Original Message----- >From: Jonathan Lang [mailto:jplang@calient.net] >Sent: Wednesday, September 19, 2001 1:38 PM >To: George Young; ccamp@ops.ietf.org >Subject: RE: Question: purpose of LMP ChannelActive message (was >relationship of RSVP-TE and LMP) > > >George, > see inline. > >Thanks, >Jonathan > >> -----Original Message----- >> From: George Young [mailto:george.young@edgeflow.com] >> Sent: Wednesday, September 19, 2001 8:42 AM >> To: Jonathan Lang; ccamp@ops.ietf.org >> Subject: RE: Question: purpose of LMP ChannelActive message (was >> relationship of RSVP-TE and LMP) >>=20 >>=20 >> Thanks for the clarification, Jonathan. >>=20 >> To see if I understand, the active call and the backup call are >> separately established by something like RSVP-TE. These two calls are >> established by PATH going from head to tail, RESV going from tail to >> head. Then for the active call ONLY, the head sends a ChannelActive >> which propagates to the tail, and then the tail sends a ChannelActive >> which propagates to the head. >>=20 >> When failure is noted by monitoring the active call, the head and the >> tail of the call are notified, and each switches over to the backup >> call. The head sends a ChannelActive message downstream and the tail >> sends a ChannelActive message upstream, and they're each=20 >propagated to >> the other end. >>=20 >> Right? >not quite right. I think you're confusing the ChannelActive=20 >message with >the Administrative Status object in GMPLS. The ChannelActive=20 >message is >used between a pair of nodes to indicate that the data channel=20 >is now active >and should be monitored. This is useful for span level protection or >between devices where GMPLS signaling (RSVP) isn't used. > >>=20 >> Regards, >> George. >> edgeflow Inc. >> 329 March Rd., Kanata, ON, Canada, K2K 2E1 >> phone: +1 613-270-9279 Ext 287 >> fax: +1 613-270-9268 >>=20 >>=20 >> >-----Original Message----- >> >From: Jonathan Lang [mailto:jplang@calient.net] >> >Sent: Tuesday, September 18, 2001 12:24 PM >> >To: George Young; ccamp@ops.ietf.org >> >Subject: RE: Question: purpose of LMP ChannelActive message (was >> >relationship of RSVP-TE and LMP) >> > >> > >> >George, >> > Sorry, I missed this one. See inline. >> > >> >-Jonathan >> > >> >> -----Original Message----- >> >> From: George Young [mailto:george.young@edgeflow.com] >> >> Sent: Tuesday, September 18, 2001 6:05 AM >> >> To: ccamp@ops.ietf.org >> >> Subject: Question: purpose of LMP ChannelActive message (was >> >> relationship of RSVP-TE and LMP) >> >>=20 >> >>=20 >> >> Didn't see any answers to this question, so I'll try to=20 >> rephrase and >> >> make it more focused. >> >>=20 >> >> What is the purpose of the LMP ChannelActive message? >> >The ChannelActive message is used to indicate that the channel=20 >> >status is >> >active and should be monitored. >> > >> >>=20 >> >> I'm assuming a call is set up through an optical cross=20 >> connect (OXC), >> >> either in response to control plane signalling (e.g. RSVP-TE) or=20 >> >> management plane input. >> >> The OXC has knowledge of which optical links are involved=20 >> >> with the call for failure monitoring purposes. >> >>=20 >> >> What additional information would the receipt of a=20 >> >> ChannelActive message convey? >> >This is useful when paths are pre-established (e.g., for path=20 >> >protection). >> >Although signaling is used to setup the resource reservation,=20 >> >on switchover >> >you need a mechanism to trigger monitoring of the data channel. >> > >> >>=20 >> >> Regards, >> >> George. >> >> edgeflow Inc. >> >> 329 March Rd., Kanata, ON, Canada, K2K 2E1 >> >> phone: +1 613-270-9279 Ext 287 >> >> fax: +1 613-270-9268 >> >>=20 >> >>=20 >> >> >-----Original Message----- >> >> >From: George Young=20 >> >> >Sent: Friday, September 07, 2001 3:47 PM >> >> >To: ccamp@ops.ietf.org >> >> >Subject: relationship of RSVP-TE and LMP >> >> > >> >> > >> >> >I have two questions about the relationship of RSVP-TE >> >> >(draft-ietf-mpls-generalized-rsvp-te-03.txt) and LMP >> >> >(draft-ietf-ccamp-lmp-00.txt). >> >> > >> >> >First of all, my understanding of the LMP ChannelActive=20 >> >> message is that >> >> >it is used to indicate to the downstream (relative to data flow >> >> >direction) node that monitoring should be used. >> >> > >> >> >The scenario I have in mind is an RSVP call setup consisting of a >> >> >bidirectional optical path. Presumably a switching node has=20 >> >> to reserve >> >> >data channels and associate them with the optical path when an=20 >> >> >RSVP PATH >> >> >message arrives. >> >> > >> >> >1) My first question is what is the timing of LMP=20 >> >> >ChannelActive messages >> >> >relative to RSVP messages. >> >> > >> >> >Do ChannelActive messages flow downstream (relative to call >> >> >establishment direction) more-or-less in phase with PATH=20 >> >> messages, and >> >> >upstream (relative to call establishment direction)=20 >> more-or-less in >> >> >phase with RESV messages? >> >> > >> >> >2) My second question is what addtional information=20 >information do >> >> >ChannelActive messages provide? >> >> > >> >> >If a link is to be monitored when associated with an=20 >> active call, it >> >> >appears the node already has this information. >> >> > >> >> >Regards, >> >> >George. >> >> >edgeflow Inc. >> >> >329 March Rd., Kanata, ON, Canada, K2K 2E1 >> >> >phone: +1 613-270-9279 Ext 287 >> >> >fax: +1 613-270-9268 >> >> > >> >> > >> >> > >> >>=20 >> > >>=20 > Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 19 Sep 2001 15:50:13 -0700 Message-ID: From: Jonathan Lang To: 'Fugui Wang' , John Drake , "'kireeti@juniper.net'" , "'yakov@juniper.net'" , "'lberger@movaz.com'" , "'dsaha@tellium.com'" , "'dbasak@accelight.com'" , "'hsandick@nortelnetworks.com'" , "'braja@tellium.com'" Cc: "'ccamp@ops.ietf.org'" Subject: RE: Questions regarding draft-ietf-ccamp-lmp-00.txt (LMP) Date: Wed, 19 Sep 2001 15:40:03 -0700 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Fugui, See inline. Thanks, Jonathan > -----Original Message----- > From: Fugui Wang [mailto:fwang@axiowave.com] > Sent: Wednesday, September 19, 2001 8:42 AM > To: Jonathan Lang; Krishna Palla; John Drake; 'kireeti@juniper.net'; > 'yakov@juniper.net'; 'lberger@movaz.com'; 'dsaha@tellium.com'; > 'dbasak@accelight.com'; 'hsandick@nortelnetworks.com'; > 'azinin@cisco.com'; 'braja@tellium.com' > Cc: 'ccamp@ops.ietf.org' > Subject: Questions regarding draft-ietf-ccamp-lmp-00.txt (LMP) > > > Hi, > > During the implementation of the LMP protocol, I have got several > questions regarding the LMP draft. Could you please clarify any of > them? > > Thanks, > Fugui > > 1. Control Channel FSM: > > (1) There are two events which could bring up the control channel, evBringUp_1a, > evBringUp_1b. Does that mean that in the MIB we need to have something like > activeUp and pasvUp in the AdminStatus of control channel table? I don't think this is necessary. > > (2) In this version, the GoingDn state is removed. When a control channel > gets evAdminDown event, it will go to Down state immediately. > Should it send a Hello or other messages with ControlChannelDown bit set to the > neighbor before it goes down? Since the Hello message is sent periodically. If we > don't send an additional Hello with ControlChannelDown bit set, most probably, the > control channel would never have a chance to send such a message. Then the remote > node will get Hello time out instead of get the event evNbrGoesDn. We will add text to the evAdminDown event indicating that Hello messages sent before going down should have the ControlChannelDown bit set. > > (3) When receiving a ConfigNack, if the Hello Config TLV is nonegotiable or > it is negotiable but we could not find the values acceptable for > both side, should we still keep sending Config message? According to the > Control Channel FSM, when getting the evConfErr, the control channel should keep in > ConfSnd state. If the configuration parameters are negotiable, and the proposed values are acceptable, a new Config message should be transmitted with these values. If the Nack'd values are non-negotiable, or if the proposed values are unacceptable, you can go into the down state (event 2) or keep trying with your values (hoping that configuration is changed at the far end). > > 2. TE link FSM: In the next version of LMP we are simplifying the TE link FSM. The TE Link can really only be in the UP, DOWN, and DEGRADED states. Link Verification can occur in either the UP or DOWN states, and could be represented using sub-states. > > (1) After verification is done (evVerDone), the FSM goes from VrfProc to > Summary state. However, if all of the Data Bearing links failed in the > verification, should the TE link still enter the Summary state and send a > LinkSummary message with no TLV? Or should it go from VrfProc to Down state > directly without sending such a meaningless LinkSummary message? (This could > happen if the Data Bearing links and TE links have not been configured in > the remote side yet and the local side has already configured and start to > verify.) > > (2) In the Summary state, the node keep sending LinkSummary message to remote > until it receives LinkSummaryAck or LinkSummaryNack. Should we have an other > events such as evSummaryTimeout so that we won't keep trying infinetely if > we couldn't get any response from the remote? > > (3) There is no event in the FSM indicating what will happen if the node > receives BeginVerify, LinkSummary. My understanding is that the receiver > side is stateless. Is that correct? > > (4) How to handle BeginVerifyNack? There is no event in the FSM > corresponding to this. > > 3. BeginVerify Message: > > The BeginVerify message brings a field VerifyInterval. But my understanding > is that the receiver side should care more about the Verify Dead > Interval than the Verify Interval since it needs to use the Verify Dead > Interval to determine when to send TestStatusFailure back. (Note that the > reciever side may not be able to get the Verify Dead Interval from the local TE link > configuration If the sender side don't know the remote TE link ID and want to learn > it through verification.) The VerifyDeadInterval is determined at the receive side and must take into account the VerifyInterval at which the Test messages are sent. Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 19 Sep 2001 11:25:12 -0700 Message-ID: <39469E08BD83D411A3D900204840EC5554C85A@vie-msgusr-01.dc.fore.com> From: "Naidu, Venkata" To: "'ccamp@ops.ietf.org'" Cc: "'mpls@uu.net'" Subject: GMPS Unnumbered Date: Wed, 19 Sep 2001 14:21:11 -0400 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" All GMPLS-Unnumbered: I have few comments on draft-ietf-ccamp-gmpls-routing-00.txt (applicable to draft-ietf-ccamp-ospf-gmpls-extensions-00.txt) * MPLS-TE extensions define Interface IP Address TLVs (Type 3 and 4) as length of 4N. Where N is the number of local IP addresses for that Interface. But in GMPLS, why can't we configure multiple "unique" Interface IDs for Unnumbered interfaces? Please allow flexibility with out restricting one ID for Unnumbered Link. There is no scarcity for local significant integers (unlike IP address) ;-) * It would be helpful, if you give more information about "why" this outgoing unnumbered link ID is chosen as "unique" in the scope of LSR. This is because, IP (Classical) has Unnumbered links defined in RFC 1812 Section 2.2.7 - which suggests to use "Router ID" (MAY be one of the Routers IP addresses) as Unnumbered Link ID. So in IP context this is not a unique ID. My question is, can we use the same IP configured unnumbered link address for GMPLS purpose? * Finally, (just for clarity) - In Routing perspective, Unnumbered Links are considered as part of GMPLS drafts. But I didn't understand why there are separate drafts draft-ietf-mpls-rsvp-unnum-02.txt and draft-ietf-mpls-crldp-unnum-02.txt in signaling perspective (more over in MPLS-WG instead of CCAMP) I think they should goto respective gmpls-rsvp and gmpls-crldp! Thanks! --Venkata Naidu Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 19 Sep 2001 10:44:56 -0700 Message-ID: From: Jonathan Lang To: 'George Young' , ccamp@ops.ietf.org Subject: RE: Question: purpose of LMP ChannelActive message (was relations hip of RSVP-TE and LMP) Date: Wed, 19 Sep 2001 10:38:19 -0700 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" George, see inline. Thanks, Jonathan > -----Original Message----- > From: George Young [mailto:george.young@edgeflow.com] > Sent: Wednesday, September 19, 2001 8:42 AM > To: Jonathan Lang; ccamp@ops.ietf.org > Subject: RE: Question: purpose of LMP ChannelActive message (was > relationship of RSVP-TE and LMP) > > > Thanks for the clarification, Jonathan. > > To see if I understand, the active call and the backup call are > separately established by something like RSVP-TE. These two calls are > established by PATH going from head to tail, RESV going from tail to > head. Then for the active call ONLY, the head sends a ChannelActive > which propagates to the tail, and then the tail sends a ChannelActive > which propagates to the head. > > When failure is noted by monitoring the active call, the head and the > tail of the call are notified, and each switches over to the backup > call. The head sends a ChannelActive message downstream and the tail > sends a ChannelActive message upstream, and they're each propagated to > the other end. > > Right? not quite right. I think you're confusing the ChannelActive message with the Administrative Status object in GMPLS. The ChannelActive message is used between a pair of nodes to indicate that the data channel is now active and should be monitored. This is useful for span level protection or between devices where GMPLS signaling (RSVP) isn't used. > > Regards, > George. > edgeflow Inc. > 329 March Rd., Kanata, ON, Canada, K2K 2E1 > phone: +1 613-270-9279 Ext 287 > fax: +1 613-270-9268 > > > >-----Original Message----- > >From: Jonathan Lang [mailto:jplang@calient.net] > >Sent: Tuesday, September 18, 2001 12:24 PM > >To: George Young; ccamp@ops.ietf.org > >Subject: RE: Question: purpose of LMP ChannelActive message (was > >relationship of RSVP-TE and LMP) > > > > > >George, > > Sorry, I missed this one. See inline. > > > >-Jonathan > > > >> -----Original Message----- > >> From: George Young [mailto:george.young@edgeflow.com] > >> Sent: Tuesday, September 18, 2001 6:05 AM > >> To: ccamp@ops.ietf.org > >> Subject: Question: purpose of LMP ChannelActive message (was > >> relationship of RSVP-TE and LMP) > >> > >> > >> Didn't see any answers to this question, so I'll try to > rephrase and > >> make it more focused. > >> > >> What is the purpose of the LMP ChannelActive message? > >The ChannelActive message is used to indicate that the channel > >status is > >active and should be monitored. > > > >> > >> I'm assuming a call is set up through an optical cross > connect (OXC), > >> either in response to control plane signalling (e.g. RSVP-TE) or > >> management plane input. > >> The OXC has knowledge of which optical links are involved > >> with the call for failure monitoring purposes. > >> > >> What additional information would the receipt of a > >> ChannelActive message convey? > >This is useful when paths are pre-established (e.g., for path > >protection). > >Although signaling is used to setup the resource reservation, > >on switchover > >you need a mechanism to trigger monitoring of the data channel. > > > >> > >> Regards, > >> George. > >> edgeflow Inc. > >> 329 March Rd., Kanata, ON, Canada, K2K 2E1 > >> phone: +1 613-270-9279 Ext 287 > >> fax: +1 613-270-9268 > >> > >> > >> >-----Original Message----- > >> >From: George Young > >> >Sent: Friday, September 07, 2001 3:47 PM > >> >To: ccamp@ops.ietf.org > >> >Subject: relationship of RSVP-TE and LMP > >> > > >> > > >> >I have two questions about the relationship of RSVP-TE > >> >(draft-ietf-mpls-generalized-rsvp-te-03.txt) and LMP > >> >(draft-ietf-ccamp-lmp-00.txt). > >> > > >> >First of all, my understanding of the LMP ChannelActive > >> message is that > >> >it is used to indicate to the downstream (relative to data flow > >> >direction) node that monitoring should be used. > >> > > >> >The scenario I have in mind is an RSVP call setup consisting of a > >> >bidirectional optical path. Presumably a switching node has > >> to reserve > >> >data channels and associate them with the optical path when an > >> >RSVP PATH > >> >message arrives. > >> > > >> >1) My first question is what is the timing of LMP > >> >ChannelActive messages > >> >relative to RSVP messages. > >> > > >> >Do ChannelActive messages flow downstream (relative to call > >> >establishment direction) more-or-less in phase with PATH > >> messages, and > >> >upstream (relative to call establishment direction) > more-or-less in > >> >phase with RESV messages? > >> > > >> >2) My second question is what addtional information information do > >> >ChannelActive messages provide? > >> > > >> >If a link is to be monitored when associated with an > active call, it > >> >appears the node already has this information. > >> > > >> >Regards, > >> >George. > >> >edgeflow Inc. > >> >329 March Rd., Kanata, ON, Canada, K2K 2E1 > >> >phone: +1 613-270-9279 Ext 287 > >> >fax: +1 613-270-9268 > >> > > >> > > >> > > >> > > > Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 19 Sep 2001 08:52:19 -0700 Message-ID: From: Fugui Wang To: "'jplang@calient.net'" , "'krishna@calient.net'" , "'jdrake@calient.net'" , "'kireeti@juniper.net'" , "'yakov@juniper.net'" , "'lberger@movaz.com'" , "'dsaha@tellium.com'" , "'dbasak@accelight.com'" , "'hsandick@nortelnetworks.com'" , "'azinin@cisco.com'" , "'braja@tellium.com'" Cc: "'ccamp@ops.ietf.org'" Subject: Questions regarding draft-ietf-ccamp-lmp-00.txt (LMP) Date: Wed, 19 Sep 2001 11:41:46 -0400 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Hi, During the implementation of the LMP protocol, I have got several questions regarding the LMP draft. Could you please clarify any of them? Thanks, Fugui 1. Control Channel FSM: (1) There are two events which could bring up the control channel, evBringUp_1a, evBringUp_1b. Does that mean that in the MIB we need to have something like activeUp and pasvUp in the AdminStatus of control channel table? (2) In this version, the GoingDn state is removed. When a control channel gets evAdminDown event, it will go to Down state immediately. Should it send a Hello or other messages with ControlChannelDown bit set to the neighbor before it goes down? Since the Hello message is sent periodically. If we don't send an additional Hello with ControlChannelDown bit set, most probably, the control channel would never have a chance to send such a message. Then the remote node will get Hello time out instead of get the event evNbrGoesDn. (3) When receiving a ConfigNack, if the Hello Config TLV is nonegotiable or it is negotiable but we could not find the values acceptable for both side, should we still keep sending Config message? According to the Control Channel FSM, when getting the evConfErr, the control channel should keep in ConfSnd state. 2. TE link FSM: (1) After verification is done (evVerDone), the FSM goes from VrfProc to Summary state. However, if all of the Data Bearing links failed in the verification, should the TE link still enter the Summary state and send a LinkSummary message with no TLV? Or should it go from VrfProc to Down state directly without sending such a meaningless LinkSummary message? (This could happen if the Data Bearing links and TE links have not been configured in the remote side yet and the local side has already configured and start to verify.) (2) In the Summary state, the node keep sending LinkSummary message to remote until it receives LinkSummaryAck or LinkSummaryNack. Should we have an other events such as evSummaryTimeout so that we won't keep trying infinetely if we couldn't get any response from the remote? (3) There is no event in the FSM indicating what will happen if the node receives BeginVerify, LinkSummary. My understanding is that the receiver side is stateless. Is that correct? (4) How to handle BeginVerifyNack? There is no event in the FSM corresponding to this. 3. BeginVerify Message: The BeginVerify message brings a field VerifyInterval. But my understanding is that the receiver side should care more about the Verify Dead Interval than the Verify Interval since it needs to use the Verify Dead Interval to determine when to send TestStatusFailure back. (Note that the reciever side may not be able to get the Verify Dead Interval from the local TE link configuration If the sender side don't know the remote TE link ID and want to learn it through verification.) Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 19 Sep 2001 08:52:12 -0700 content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Subject: RE: Question: purpose of LMP ChannelActive message (was relationship of RSVP-TE and LMP) Date: Wed, 19 Sep 2001 11:41:30 -0400 Message-ID: Thread-Topic: Question: purpose of LMP ChannelActive message (was relationship of RSVP-TE and LMP) Thread-Index: AcFAXtIvE2orz7bLTquxUOwZ8IsqOQAwmwhQ From: "George Young" To: "Jonathan Lang" , Thanks for the clarification, Jonathan. To see if I understand, the active call and the backup call are separately established by something like RSVP-TE. These two calls are established by PATH going from head to tail, RESV going from tail to head. Then for the active call ONLY, the head sends a ChannelActive which propagates to the tail, and then the tail sends a ChannelActive which propagates to the head. When failure is noted by monitoring the active call, the head and the tail of the call are notified, and each switches over to the backup call. The head sends a ChannelActive message downstream and the tail sends a ChannelActive message upstream, and they're each propagated to the other end. Right? Regards, George. edgeflow Inc. 329 March Rd., Kanata, ON, Canada, K2K 2E1 phone: +1 613-270-9279 Ext 287 fax: +1 613-270-9268 >-----Original Message----- >From: Jonathan Lang [mailto:jplang@calient.net] >Sent: Tuesday, September 18, 2001 12:24 PM >To: George Young; ccamp@ops.ietf.org >Subject: RE: Question: purpose of LMP ChannelActive message (was >relationship of RSVP-TE and LMP) > > >George, > Sorry, I missed this one. See inline. > >-Jonathan > >> -----Original Message----- >> From: George Young [mailto:george.young@edgeflow.com] >> Sent: Tuesday, September 18, 2001 6:05 AM >> To: ccamp@ops.ietf.org >> Subject: Question: purpose of LMP ChannelActive message (was >> relationship of RSVP-TE and LMP) >>=20 >>=20 >> Didn't see any answers to this question, so I'll try to rephrase and >> make it more focused. >>=20 >> What is the purpose of the LMP ChannelActive message? >The ChannelActive message is used to indicate that the channel=20 >status is >active and should be monitored. > >>=20 >> I'm assuming a call is set up through an optical cross connect (OXC), >> either in response to control plane signalling (e.g. RSVP-TE) or=20 >> management plane input. >> The OXC has knowledge of which optical links are involved=20 >> with the call for failure monitoring purposes. >>=20 >> What additional information would the receipt of a=20 >> ChannelActive message convey? >This is useful when paths are pre-established (e.g., for path=20 >protection). >Although signaling is used to setup the resource reservation,=20 >on switchover >you need a mechanism to trigger monitoring of the data channel. > >>=20 >> Regards, >> George. >> edgeflow Inc. >> 329 March Rd., Kanata, ON, Canada, K2K 2E1 >> phone: +1 613-270-9279 Ext 287 >> fax: +1 613-270-9268 >>=20 >>=20 >> >-----Original Message----- >> >From: George Young=20 >> >Sent: Friday, September 07, 2001 3:47 PM >> >To: ccamp@ops.ietf.org >> >Subject: relationship of RSVP-TE and LMP >> > >> > >> >I have two questions about the relationship of RSVP-TE >> >(draft-ietf-mpls-generalized-rsvp-te-03.txt) and LMP >> >(draft-ietf-ccamp-lmp-00.txt). >> > >> >First of all, my understanding of the LMP ChannelActive=20 >> message is that >> >it is used to indicate to the downstream (relative to data flow >> >direction) node that monitoring should be used. >> > >> >The scenario I have in mind is an RSVP call setup consisting of a >> >bidirectional optical path. Presumably a switching node has=20 >> to reserve >> >data channels and associate them with the optical path when an=20 >> >RSVP PATH >> >message arrives. >> > >> >1) My first question is what is the timing of LMP=20 >> >ChannelActive messages >> >relative to RSVP messages. >> > >> >Do ChannelActive messages flow downstream (relative to call >> >establishment direction) more-or-less in phase with PATH=20 >> messages, and >> >upstream (relative to call establishment direction) more-or-less in >> >phase with RESV messages? >> > >> >2) My second question is what addtional information information do >> >ChannelActive messages provide? >> > >> >If a link is to be monitored when associated with an active call, it >> >appears the node already has this information. >> > >> >Regards, >> >George. >> >edgeflow Inc. >> >329 March Rd., Kanata, ON, Canada, K2K 2E1 >> >phone: +1 613-270-9279 Ext 287 >> >fax: +1 613-270-9268 >> > >> > >> > >>=20 > Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 18 Sep 2001 09:32:51 -0700 Message-ID: From: Jonathan Lang To: 'George Young' , ccamp@ops.ietf.org Subject: RE: Question: purpose of LMP ChannelActive message (was relations hip of RSVP-TE and LMP) Date: Tue, 18 Sep 2001 09:23:51 -0700 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" George, Sorry, I missed this one. See inline. -Jonathan > -----Original Message----- > From: George Young [mailto:george.young@edgeflow.com] > Sent: Tuesday, September 18, 2001 6:05 AM > To: ccamp@ops.ietf.org > Subject: Question: purpose of LMP ChannelActive message (was > relationship of RSVP-TE and LMP) > > > Didn't see any answers to this question, so I'll try to rephrase and > make it more focused. > > What is the purpose of the LMP ChannelActive message? The ChannelActive message is used to indicate that the channel status is active and should be monitored. > > I'm assuming a call is set up through an optical cross connect (OXC), > either in response to control plane signalling (e.g. RSVP-TE) or > management plane input. > The OXC has knowledge of which optical links are involved > with the call for failure monitoring purposes. > > What additional information would the receipt of a > ChannelActive message convey? This is useful when paths are pre-established (e.g., for path protection). Although signaling is used to setup the resource reservation, on switchover you need a mechanism to trigger monitoring of the data channel. > > Regards, > George. > edgeflow Inc. > 329 March Rd., Kanata, ON, Canada, K2K 2E1 > phone: +1 613-270-9279 Ext 287 > fax: +1 613-270-9268 > > > >-----Original Message----- > >From: George Young > >Sent: Friday, September 07, 2001 3:47 PM > >To: ccamp@ops.ietf.org > >Subject: relationship of RSVP-TE and LMP > > > > > >I have two questions about the relationship of RSVP-TE > >(draft-ietf-mpls-generalized-rsvp-te-03.txt) and LMP > >(draft-ietf-ccamp-lmp-00.txt). > > > >First of all, my understanding of the LMP ChannelActive > message is that > >it is used to indicate to the downstream (relative to data flow > >direction) node that monitoring should be used. > > > >The scenario I have in mind is an RSVP call setup consisting of a > >bidirectional optical path. Presumably a switching node has > to reserve > >data channels and associate them with the optical path when an > >RSVP PATH > >message arrives. > > > >1) My first question is what is the timing of LMP > >ChannelActive messages > >relative to RSVP messages. > > > >Do ChannelActive messages flow downstream (relative to call > >establishment direction) more-or-less in phase with PATH > messages, and > >upstream (relative to call establishment direction) more-or-less in > >phase with RESV messages? > > > >2) My second question is what addtional information information do > >ChannelActive messages provide? > > > >If a link is to be monitored when associated with an active call, it > >appears the node already has this information. > > > >Regards, > >George. > >edgeflow Inc. > >329 March Rd., Kanata, ON, Canada, K2K 2E1 > >phone: +1 613-270-9279 Ext 287 > >fax: +1 613-270-9268 > > > > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 18 Sep 2001 06:09:26 -0700 content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Subject: Question: purpose of LMP ChannelActive message (was relationship of RSVP-TE and LMP) Date: Tue, 18 Sep 2001 09:05:25 -0400 Message-ID: Thread-Topic: relationship of RSVP-TE and LMP Thread-Index: AcE31ey3QHH4It+VQsGgnX9bKulvYwIa3sKA From: "George Young" To: Didn't see any answers to this question, so I'll try to rephrase and make it more focused. What is the purpose of the LMP ChannelActive message? I'm assuming a call is set up through an optical cross connect (OXC), either in response to control plane signalling (e.g. RSVP-TE) or management plane input. The OXC has knowledge of which optical links are involved with the call for failure monitoring purposes. What additional information would the receipt of a ChannelActive message convey? Regards, George. edgeflow Inc. 329 March Rd., Kanata, ON, Canada, K2K 2E1 phone: +1 613-270-9279 Ext 287 fax: +1 613-270-9268 >-----Original Message----- >From: George Young=20 >Sent: Friday, September 07, 2001 3:47 PM >To: ccamp@ops.ietf.org >Subject: relationship of RSVP-TE and LMP > > >I have two questions about the relationship of RSVP-TE >(draft-ietf-mpls-generalized-rsvp-te-03.txt) and LMP >(draft-ietf-ccamp-lmp-00.txt). > >First of all, my understanding of the LMP ChannelActive message is that >it is used to indicate to the downstream (relative to data flow >direction) node that monitoring should be used. > >The scenario I have in mind is an RSVP call setup consisting of a >bidirectional optical path. Presumably a switching node has to reserve >data channels and associate them with the optical path when an=20 >RSVP PATH >message arrives. > >1) My first question is what is the timing of LMP=20 >ChannelActive messages >relative to RSVP messages. > >Do ChannelActive messages flow downstream (relative to call >establishment direction) more-or-less in phase with PATH messages, and >upstream (relative to call establishment direction) more-or-less in >phase with RESV messages? > >2) My second question is what addtional information information do >ChannelActive messages provide? > >If a link is to be monitored when associated with an active call, it >appears the node already has this information. > >Regards, >George. >edgeflow Inc. >329 March Rd., Kanata, ON, Canada, K2K 2E1 >phone: +1 613-270-9279 Ext 287 >fax: +1 613-270-9268 > > > Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 17 Sep 2001 16:17:19 -0700 Message-Id: <200106261102.HAA27267@ietf.org> Mime-Version: 1.0 Content-Type: Multipart/Mixed; Boundary="NextPart" To: IETF-Announce: ; CC: ccamp@ops.ietf.org From: Internet-Drafts@ietf.org Reply-to: Internet-Drafts@ietf.org Subject: I-D ACTION:draft-srisuresh-ospf-te-00.txt Date: Tue, 26 Jun 2001 07:02:32 -0400 --NextPart A New Internet-Draft is available from the on-line Internet-Drafts directories. Title : New TE LSAs to extend OSPF for Traffic Engineering Author(s) : P. Srisuresh, P. Joseph Filename : draft-srisuresh-ospf-te-00.txt Pages : 36 Date : 25-Jun-01 OSPF is a well established link state routing protocol used for topology discovery and computing forwarding table based on shortest-Path criteria. Traffic Engineering extensions (OSPF-TE) will use criteria different from shortest-path so as to route traffic around congestion paths and meet varying Service Level agreements. A URL for this Internet-Draft is: http://www.ietf.org/internet-drafts/draft-srisuresh-ospf-te-00.txt Internet-Drafts are also available by anonymous FTP. Login with the username "anonymous" and a password of your e-mail address. After logging in, type "cd internet-drafts" and then "get draft-srisuresh-ospf-te-00.txt". A list of Internet-Drafts directories can be found in http://www.ietf.org/shadow.html or ftp://ftp.ietf.org/ietf/1shadow-sites.txt Internet-Drafts can also be obtained by e-mail. Send a message to: mailserv@ietf.org. In the body type: "FILE /internet-drafts/draft-srisuresh-ospf-te-00.txt". NOTE: The mail server at ietf.org can return the document in MIME-encoded form by using the "mpack" utility. To use this feature, insert the command "ENCODING mime" before the "FILE" command. To decode the response(s), you will need "munpack" or a MIME-compliant mail reader. Different MIME-compliant mail readers exhibit different behavior, especially when dealing with "multipart" MIME messages (i.e. documents which have been split up into multiple messages), so check your local documentation on how to manipulate these messages. Below is the data which will enable a MIME compliant mail reader implementation to automatically retrieve the ASCII version of the Internet-Draft. --NextPart Content-Type: Multipart/Alternative; Boundary="OtherAccess" --OtherAccess Content-Type: Message/External-body; access-type="mail-server"; server="mailserv@ietf.org" Content-Type: text/plain Content-ID: <20010625101030.I-D@ietf.org> ENCODING mime FILE /internet-drafts/draft-srisuresh-ospf-te-00.txt --OtherAccess Content-Type: Message/External-body; name="draft-srisuresh-ospf-te-00.txt"; site="ftp.ietf.org"; access-type="anon-ftp"; directory="internet-drafts" Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 17 Sep 2001 16:17:16 -0700 Message-Id: <200107061050.GAA25924@ietf.org> Mime-Version: 1.0 Content-Type: Multipart/Mixed; Boundary="NextPart" To: IETF-Announce: ; CC: mpls@uu.net, ccamp@ops.ietf.org From: Internet-Drafts@ietf.org Reply-to: Internet-Drafts@ietf.org Subject: I-D ACTION:draft-iwata-mpls-crankback-01.txt Date: Fri, 06 Jul 2001 06:50:43 -0400 --NextPart A New Internet-Draft is available from the on-line Internet-Drafts directories. Title : Crankback Routing Extensions for MPLS Signaling Author(s) : A. Iwata, N. Fujita, G. Ash, A. Farrel Filename : draft-iwata-mpls-crankback-01.txt Pages : 30 Date : 05-Jul-01 This draft proposes crankback routing extensions for CR-LDP signaling and for RSVP-TE signaling. Recently, several routing protocol extensions for advertising resource information in addition to topology information have been proposed for use in distributed constraint-based routing. In such a distributed routing environment, however, the information used to compute a constraint-based path may be out of date. This means that LSP setup requests may be blocked by links or nodes without sufficient resources. This draft specifies crankback routing extensions for CR-LDP and RSVP-TE so that the label request can be retried on an alternate path that detours around the blocked link or node upon a setup failure. Furthermore, the crankback routing schemes can also be applied to LSP restoration by indicating the location of the failure link or node. This would significantly improve the successful recovery ratio for failed LSPs, especially in situations where a large number of setup requests are triggered at the same time. A URL for this Internet-Draft is: http://www.ietf.org/internet-drafts/draft-iwata-mpls-crankback-01.txt Internet-Drafts are also available by anonymous FTP. Login with the username "anonymous" and a password of your e-mail address. After logging in, type "cd internet-drafts" and then "get draft-iwata-mpls-crankback-01.txt". A list of Internet-Drafts directories can be found in http://www.ietf.org/shadow.html or ftp://ftp.ietf.org/ietf/1shadow-sites.txt Internet-Drafts can also be obtained by e-mail. Send a message to: mailserv@ietf.org. In the body type: "FILE /internet-drafts/draft-iwata-mpls-crankback-01.txt". NOTE: The mail server at ietf.org can return the document in MIME-encoded form by using the "mpack" utility. To use this feature, insert the command "ENCODING mime" before the "FILE" command. To decode the response(s), you will need "munpack" or a MIME-compliant mail reader. Different MIME-compliant mail readers exhibit different behavior, especially when dealing with "multipart" MIME messages (i.e. documents which have been split up into multiple messages), so check your local documentation on how to manipulate these messages. Below is the data which will enable a MIME compliant mail reader implementation to automatically retrieve the ASCII version of the Internet-Draft. --NextPart Content-Type: Multipart/Alternative; Boundary="OtherAccess" --OtherAccess Content-Type: Message/External-body; access-type="mail-server"; server="mailserv@ietf.org" Content-Type: text/plain Content-ID: <20010705113912.I-D@ietf.org> ENCODING mime FILE /internet-drafts/draft-iwata-mpls-crankback-01.txt --OtherAccess Content-Type: Message/External-body; name="draft-iwata-mpls-crankback-01.txt"; site="ftp.ietf.org"; access-type="anon-ftp"; directory="internet-d Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 13 Sep 2001 13:14:09 -0700 Message-ID: <9D6D37E97A57D411BB7C00508BAE29C901AF26FC@main.mahinetworks.com> From: Abhimanyu Das To: 'Edward Harrison' Cc: "'ccamp@ops.ietf.org'" Subject: RE: FW: Questions on draft-ietf-mpls-bundle-00 Date: Thu, 13 Sep 2001 13:11:55 -0700 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Hi, I agree with Ed that it does make sense to keep both the COMPONENT_IF_UPSTREAM and COMPONENT_IF_DOWNSTREAM objects..since in bundled links, one should allow for the flexibility of using different interfaces for upstream and downstream LSPs. In fact, it seems bundling is the only case where you can allow for this flexibility of choosing different interfaces for upstream and downstream LSPs because you have a choice of selecting a transmitter on one link and a receiver on another link within the same bundle. In non-bundled cases(numbered or unnumbered), the link receiver and transmitter is combined within a single interface..so one seems forced in practice to use the same interface for both upstream and downstream directions. Abhimanyu > -----Original Message----- > From: Edward Harrison [mailto:eph@dataconnection.com] > Sent: Thursday, September 13, 2001 2:56 AM > To: 'Yakov Rekhter' > Cc: Nic Larkin; 'ccamp@ops.ietf.org' > Subject: RE: FW: Questions on draft-ietf-mpls-bundle-00 > > > Yakov, > > Can you clarify whether you are proposing removing the > COMPONENT_IF_UPSTREAM > TLV from the INTERFACE_ID hop object (or were you just saying > that Steve > needn't use it if he doesn't want to)? > > As far as I can see, there is no restriction in LMP, for example, that > requires data links to be grouped into upstream and > downstream pairs. The > ability to signal separate upstream and downstream links, > therefore seems to > be potentially useful. > > Thanks, > > Ed Harrison > > ----------------- > Edward Harrison mailto:eph@dataconnection.com > Network Convergence Group > Data Connection Ltd. > http://www.dataconnection.com/ > Tel: +44 20 8366 1177 Fax: +44 20 8363 1468 > > > -----Original Message----- > From: Yakov Rekhter [mailto:yakov@juniper.net] > Sent: 12 September 2001 20:56 > To: Atkinson, Stephen > Cc: 'ccamp@ops.ietf.org' > Subject: Re: FW: Questions on draft-ietf-mpls-bundle-00 > > > Stephen, > > > I didn't get any comments on this. Would anyone care to > > respond? > > Sorry for the delay... comments below.... > > > Thanks, > > steve atkinson > > > > > -----Original Message----- > > > From: Atkinson, Stephen > > > Sent: Monday, September 10, 2001 4:11 PM > > > To: ccamp@ops.ietf.org > > > Subject: Questions on draft-ietf-mpls-bundle-00 > > > > > > I've been reading the draft-ietf-mpls-bundle-00 spec and have > > > a couple questions. > > > > > > Basically I'm trying to focus on why signaling unnumbered > > > components using > > > Explicit Indication by Interface ID is different from all > > > other cases. This being the > > > only case which specifies an interface (component) for each > > > direction of a bi-directional LSP. > > My suggestion would be to make signaling unnumbered components > using Explicit Indication by Interface ID the same as all other > cases. That is, when signaling unnumbered components using > Explicit Indication the signaling would carry just *one* > Interface ID, and this Interface ID would identify the component > link that should be used in *both* directions. > > Yakov. > Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 13 Sep 2001 09:40:19 -0700 Message-ID: <87009604743AD411B1F600508BA0F959632450@XOVER.dedham.mindspeed.com> From: "Atkinson, Stephen" To: "'ccamp@ops.ietf.org'" Cc: 'Edward Harrison' Subject: RE: FW: Questions on draft-ietf-mpls-bundle-00 Date: Thu, 13 Sep 2001 12:39:05 -0400 MIME-Version: 1.0 Content-Type: text/plain Hi Ed, I agree, comments below. Regards, steve atkinson > -----Original Message----- > From: Edward Harrison [mailto:eph@dataconnection.com] > Sent: Thursday, September 13, 2001 11:06 AM > To: 'Yakov Rekhter' > Cc: Nic Larkin; 'ccamp@ops.ietf.org' > Subject: RE: FW: Questions on draft-ietf-mpls-bundle-00 > > > Yakov, > > > > Can you clarify whether you are proposing removing the > COMPONENT_IF_UPSTREAM > > > TLV from the INTERFACE_ID hop object (or were you just > saying that Steve > > > needn't use it if he doesn't want to)? > > > > I am proposing to remove the COMPONENT_IF_UPSTREAM. > > In this case, am I right in thinking it makes sense to remove > COMPONENT_IF_DOWNSTREAM too, and signal the (single) > component interface > identifier, in the IPv4, IPv6 or IF_INDEX TLV? This is how I felt it should work. I don't see any additional value in COMPONENT_IF_DOWNSTREAM that can't be handled by the IF_INDEX TLV. Both carry the same information, the unnumbered interface identifier. > > We'd be interested to hear if anyone on the mailing list > finds this approach > overly restrictive. > > Ed. > > -----Original Message----- > From: Yakov Rekhter [mailto:yakov@juniper.net] > Sent: 13 September 2001 14:58 > To: Edward Harrison > Cc: 'Yakov Rekhter'; Nic Larkin; 'ccamp@ops.ietf.org' > Subject: Re: FW: Questions on draft-ietf-mpls-bundle-00 > > > Ed, > > > Can you clarify whether you are proposing removing the > COMPONENT_IF_UPSTREAM > > TLV from the INTERFACE_ID hop object (or were you just > saying that Steve > > needn't use it if he doesn't want to)? > > I am proposing to remove the COMPONENT_IF_UPSTREAM. > > Yakov. > Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 13 Sep 2001 08:10:21 -0700 Message-ID: <53F74F5A7B94D511841C00B0D0AB16F8135ECE@baker.datcon.co.uk> From: Edward Harrison To: 'Yakov Rekhter' Cc: Nic Larkin , "'ccamp@ops.ietf.org'" Subject: RE: FW: Questions on draft-ietf-mpls-bundle-00 Date: Thu, 13 Sep 2001 16:06:19 +0100 Deferred-Delivery: Thu, 13 Sep 2001 16:08:00 +0100 MIME-Version: 1.0 Content-Type: text/plain Yakov, > > Can you clarify whether you are proposing removing the COMPONENT_IF_UPSTREAM > > TLV from the INTERFACE_ID hop object (or were you just saying that Steve > > needn't use it if he doesn't want to)? > > I am proposing to remove the COMPONENT_IF_UPSTREAM. In this case, am I right in thinking it makes sense to remove COMPONENT_IF_DOWNSTREAM too, and signal the (single) component interface identifier, in the IPv4, IPv6 or IF_INDEX TLV? We'd be interested to hear if anyone on the mailing list finds this approach overly restrictive. Ed. -----Original Message----- From: Yakov Rekhter [mailto:yakov@juniper.net] Sent: 13 September 2001 14:58 To: Edward Harrison Cc: 'Yakov Rekhter'; Nic Larkin; 'ccamp@ops.ietf.org' Subject: Re: FW: Questions on draft-ietf-mpls-bundle-00 Ed, > Can you clarify whether you are proposing removing the COMPONENT_IF_UPSTREAM > TLV from the INTERFACE_ID hop object (or were you just saying that Steve > needn't use it if he doesn't want to)? I am proposing to remove the COMPONENT_IF_UPSTREAM. Yakov. Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 13 Sep 2001 07:00:39 -0700 Message-Id: <200109131358.f8DDwS073949@merlot.juniper.net> To: Edward Harrison cc: "'Yakov Rekhter'" , Nic Larkin , "'ccamp@ops.ietf.org'" Subject: Re: FW: Questions on draft-ietf-mpls-bundle-00 MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-ID: <22664.1000389508.1@juniper.net> Date: Thu, 13 Sep 2001 06:58:28 -0700 From: Yakov Rekhter Ed, > Can you clarify whether you are proposing removing the COMPONENT_IF_UPSTREAM > TLV from the INTERFACE_ID hop object (or were you just saying that Steve > needn't use it if he doesn't want to)? I am proposing to remove the COMPONENT_IF_UPSTREAM. Yakov. Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 13 Sep 2001 02:59:09 -0700 Message-ID: <53F74F5A7B94D511841C00B0D0AB16F8135EC6@baker.datcon.co.uk> From: Edward Harrison To: 'Yakov Rekhter' Cc: Nic Larkin , "'ccamp@ops.ietf.org'" Subject: RE: FW: Questions on draft-ietf-mpls-bundle-00 Date: Thu, 13 Sep 2001 10:55:41 +0100 Deferred-Delivery: Thu, 13 Sep 2001 10:57:00 +0100 MIME-Version: 1.0 Content-Type: text/plain Yakov, Can you clarify whether you are proposing removing the COMPONENT_IF_UPSTREAM TLV from the INTERFACE_ID hop object (or were you just saying that Steve needn't use it if he doesn't want to)? As far as I can see, there is no restriction in LMP, for example, that requires data links to be grouped into upstream and downstream pairs. The ability to signal separate upstream and downstream links, therefore seems to be potentially useful. Thanks, Ed Harrison ----------------- Edward Harrison mailto:eph@dataconnection.com Network Convergence Group Data Connection Ltd. http://www.dataconnection.com/ Tel: +44 20 8366 1177 Fax: +44 20 8363 1468 -----Original Message----- From: Yakov Rekhter [mailto:yakov@juniper.net] Sent: 12 September 2001 20:56 To: Atkinson, Stephen Cc: 'ccamp@ops.ietf.org' Subject: Re: FW: Questions on draft-ietf-mpls-bundle-00 Stephen, > I didn't get any comments on this. Would anyone care to > respond? Sorry for the delay... comments below.... > Thanks, > steve atkinson > > > -----Original Message----- > > From: Atkinson, Stephen > > Sent: Monday, September 10, 2001 4:11 PM > > To: ccamp@ops.ietf.org > > Subject: Questions on draft-ietf-mpls-bundle-00 > > > > I've been reading the draft-ietf-mpls-bundle-00 spec and have > > a couple questions. > > > > Basically I'm trying to focus on why signaling unnumbered > > components using > > Explicit Indication by Interface ID is different from all > > other cases. This being the > > only case which specifies an interface (component) for each > > direction of a bi-directional LSP. My suggestion would be to make signaling unnumbered components using Explicit Indication by Interface ID the same as all other cases. That is, when signaling unnumbered components using Explicit Indication the signaling would carry just *one* Interface ID, and this Interface ID would identify the component link that should be used in *both* directions. Yakov. Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 12 Sep 2001 13:01:28 -0700 Message-Id: <200109121956.f8CJu0027557@merlot.juniper.net> To: "Atkinson, Stephen" cc: "'ccamp@ops.ietf.org'" Subject: Re: FW: Questions on draft-ietf-mpls-bundle-00 MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-ID: <6210.1000324560.1@juniper.net> Date: Wed, 12 Sep 2001 12:56:00 -0700 From: Yakov Rekhter Stephen, > I didn't get any comments on this. Would anyone care to > respond? Sorry for the delay... comments below.... > Thanks, > steve atkinson > > > -----Original Message----- > > From: Atkinson, Stephen > > Sent: Monday, September 10, 2001 4:11 PM > > To: ccamp@ops.ietf.org > > Subject: Questions on draft-ietf-mpls-bundle-00 > > > > I've been reading the draft-ietf-mpls-bundle-00 spec and have > > a couple questions. > > > > Basically I'm trying to focus on why signaling unnumbered > > components using > > Explicit Indication by Interface ID is different from all > > other cases. This being the > > only case which specifies an interface (component) for each > > direction of a bi-directional LSP. My suggestion would be to make signaling unnumbered components using Explicit Indication by Interface ID the same as all other cases. That is, when signaling unnumbered components using Explicit Indication the signaling would carry just *one* Interface ID, and this Interface ID would identify the component link that should be used in *both* directions. Yakov. Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 12 Sep 2001 12:22:56 -0700 Message-ID: <87009604743AD411B1F600508BA0F95963244F@XOVER.dedham.mindspeed.com> From: "Atkinson, Stephen" To: "'ccamp@ops.ietf.org'" Subject: FW: Questions on draft-ietf-mpls-bundle-00 Date: Wed, 12 Sep 2001 15:17:15 -0400 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" I didn't get any comments on this. Would anyone care to respond? Thanks, steve atkinson > -----Original Message----- > From: Atkinson, Stephen > Sent: Monday, September 10, 2001 4:11 PM > To: ccamp@ops.ietf.org > Subject: Questions on draft-ietf-mpls-bundle-00 > > I've been reading the draft-ietf-mpls-bundle-00 spec and have > a couple questions. > > Basically I'm trying to focus on why signaling unnumbered > components using > Explicit Indication by Interface ID is different from all > other cases. This being the > only case which specifies an interface (component) for each > direction of a > bi-directional LSP. > > Bundling section 3.3.2 (Explicit Indication) requires that > the upstream node specify > the component interface to use in each direction when setting > up bi-directional > LSPs using unnumbered components. > > Bundling section 3.3.1 (Implicit Indication) doesn't have > this requirement. Apparently > when signaling using "Implicit Indication" (for a > bi-directional LSP) the reverse direction > can be inferred. Is this true? If so is this behavior > specified anywhere? > > Also, when signaling a bi-directional LSP using numbered > components the > reverse direction component can apparently be inferred using > Implicit Indication or > Explicit Indication by Interface ID. Is this true? > > Assuming the statements above are correct. Why must a > bi-directional LSP being setup using > Explicit Indication and unnumbered components, specify a > component for each direction, while > each of the following only requires specifying a single > component (or interface)? > > - unnumbered component signaled implicitly > - numbered component signaled implicitly > - numbered component signaled explicitly > - unnumbered interface (not bundling) signaled implicitly > or explicitly > - numbered interface (not bundling) signaled implicitly or > explicitly > > > Thanks, > steve atkinson > > > Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 10 Sep 2001 18:40:20 -0700 Message-Id: <200109110139.f8B1d0032976@merlot.juniper.net> To: =?iso-2022-jp?B?VG9tb2hpa28bJEIhIRsoQllBR1lV?= cc: ccamp@ops.ietf.org Subject: Re: Q about draft-ccamp-gmpls-routing-00 and ospf-gmpls-extensions-00 MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-ID: <48767.1000172340.1@juniper.net> Date: Mon, 10 Sep 2001 18:39:00 -0700 From: Yakov Rekhter Yagyu, > I have a question about draft-ietf-ccamp-gmpls-routing-00 and > draft-ietf-ccamp-ospf-gmpls-extensions-00. > > In 6.4.4 of routing draft, when Switching capability is LSC, > > The additional information includes Reservable Bandwidth per > priority, which specifies the bandwidth of an LSP that could be > supported by the interface at a given priority number. > > However, In 5.6 of ospf-extensions darft, > > When the Switching Capability field is LSC, there is no specific > information. > > There is a confilct, which is true? In the case of LSC the Reservable Bandwidth per priority is encoded in Max LSP of the Interface Switching Capability Descriptor. Yakov. Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 10 Sep 2001 18:18:11 -0700 Message-ID: <005801c13a5f$41b0e260$c50510ac@netlab.nec.co.jp> Reply-To: =?iso-2022-jp?B?VG9tb2hpa28bJEIhIRsoQllBR1lV?= From: =?iso-2022-jp?B?VG9tb2hpa28bJEIhIRsoQllBR1lV?= To: Subject: Q about draft-ccamp-gmpls-routing-00 and ospf-gmpls-extensions-00 Date: Tue, 11 Sep 2001 10:15:33 +0900 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-2022-jp" Content-Transfer-Encoding: 7bit I have a question about draft-ietf-ccamp-gmpls-routing-00 and draft-ietf-ccamp-ospf-gmpls-extensions-00. In 6.4.4 of routing draft, when Switching capability is LSC, The additional information includes Reservable Bandwidth per priority, which specifies the bandwidth of an LSP that could be supported by the interface at a given priority number. However, In 5.6 of ospf-extensions darft, When the Switching Capability field is LSC, there is no specific information. There is a confilct, which is true? In my personal opinion, I prefer ospf-extensions draft's method. Thanks. Yagyu@NEC Corporaiton Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 10 Sep 2001 13:13:05 -0700 Message-ID: <87009604743AD411B1F600508BA0F95963244D@XOVER.dedham.mindspeed.com> From: "Atkinson, Stephen" To: ccamp@ops.ietf.org Subject: Questions on draft-ietf-mpls-bundle-00 Date: Mon, 10 Sep 2001 16:10:47 -0400 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" I've been reading the draft-ietf-mpls-bundle-00 spec and have a couple questions. Basically I'm trying to focus on why signaling unnumbered components using Explicit Indication by Interface ID is different from all other cases. This being the only case which specifies an interface (component) for each direction of a bi-directional LSP. Bundling section 3.3.2 (Explicit Indication) requires that the upstream node specify the component interface to use in each direction when setting up bi-directional LSPs using unnumbered components. Bundling section 3.3.1 (Implicit Indication) doesn't have this requirement. Apparently when signaling using "Implicit Indication" (for a bi-directional LSP) the reverse direction can be inferred. Is this true? If so is this behavior specified anywhere? Also, when signaling a bi-directional LSP using numbered components the reverse direction component can apparently be inferred using Implicit Indication or Explicit Indication by Interface ID. Is this true? Assuming the statements above are correct. Why must a bi-directional LSP being setup using Explicit Indication and unnumbered components, specify a component for each direction, while each of the following only requires specifying a single component (or interface)? - unnumbered component signaled implicitly - numbered component signaled implicitly - numbered component signaled explicitly - unnumbered interface (not bundling) signaled implicitly or explicitly - numbered interface (not bundling) signaled implicitly or explicitly Thanks, steve atkinson Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 10 Sep 2001 10:55:53 -0700 Message-ID: From: Jonathan Lang To: 'George Young' , Jonathan Lang Cc: ccamp@ops.ietf.org Subject: RE: relationship of ChannelFail and Data Link FSM Date: Mon, 10 Sep 2001 10:50:32 -0700 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" George, Yes. thanks, Jonathan > -----Original Message----- > From: George Young [mailto:george.young@edgeflow.com] > Sent: Monday, September 10, 2001 10:52 AM > To: Jonathan Lang > Cc: ccamp@ops.ietf.org > Subject: RE: relationship of ChannelFail and Data Link FSM > > > Hello Jonathan, > > Thanks for the clarification. > > Does this mean there should be a transition in the Active > Data Link FSM > Description from Up/Alloc to Down, triggered by a 'failure localized' > event? > > Regards, > George. > edgeflow Inc. > 329 March Rd., Kanata, ON, Canada, K2K 2E1 > phone: +1 613-270-9279 Ext 287 > fax: +1 613-270-9268 > > > >-----Original Message----- > >From: Jonathan Lang [mailto:jplang@calient.net] > >Sent: Monday, September 10, 2001 1:13 PM > >To: George Young; ccamp@ops.ietf.org > >Subject: RE: relationship of ChannelFail and Data Link FSM > > > > > >George, > > The upstream node should put the data link into the Down > state. The > >downstream node remains in the Up/Alloc state since it doesn't > >know if the > >failure is with this particular link or if it is further > upstream. The > >reason to remain in up/alloc is that it may see restored > >traffic over the > >link if the failure is repaired upstream. > > > >Thanks, > >Jonathan > > > >> -----Original Message----- > >> From: George Young [mailto:george.young@edgeflow.com] > >> Sent: Friday, September 07, 2001 10:13 AM > >> To: ccamp@ops.ietf.org > >> Subject: relationship of ChannelFail and Data Link FSM > >> > >> > >> Regarding ID draft-ietf-ccamp-lmp-00.txt, I have a > question about the > >> ChannelFail message and the Data Link FSM. > >> > >> Firstly, my understanding is that when channel monitoring detects a > >> failure, an LMP ChannelFail message is sent upstream, > relative to the > >> direction of transmission. A node which receives this message > >> and knows > >> via monitoring that its upstream link is ok has localized > >the failure. > >> > >> My question is does this cause any state change in the > Data Link FSM, > >> and if so, does it occur only in the node which localizes > >the failure, > >> or all nodes which are aware of the failure. The current > FSM does not > >> seem to address this. > >> > >> Regards, > >> George. > >> edgeflow Inc. > >> 329 March Rd., Kanata, ON, Canada, K2K 2E1 > >> phone: +1 613-270-9279 Ext 287 > >> fax: +1 613-270-9268 > >> > >> > > > Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 10 Sep 2001 10:54:28 -0700 content-class: urn:content-classes:message Subject: RE: relationship of ChannelFail and Data Link FSM MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Date: Mon, 10 Sep 2001 13:51:35 -0400 Message-ID: Thread-Topic: relationship of ChannelFail and Data Link FSM Thread-Index: AcE6HE6sDKcTWXYVSzuPXTjrDQYSBwABAI1A From: "George Young" To: "Jonathan Lang" Cc: Hello Jonathan, Thanks for the clarification. Does this mean there should be a transition in the Active Data Link FSM Description from Up/Alloc to Down, triggered by a 'failure localized' event? Regards, George. edgeflow Inc. 329 March Rd., Kanata, ON, Canada, K2K 2E1 phone: +1 613-270-9279 Ext 287 fax: +1 613-270-9268 >-----Original Message----- >From: Jonathan Lang [mailto:jplang@calient.net] >Sent: Monday, September 10, 2001 1:13 PM >To: George Young; ccamp@ops.ietf.org >Subject: RE: relationship of ChannelFail and Data Link FSM > > >George, > The upstream node should put the data link into the Down state. The >downstream node remains in the Up/Alloc state since it doesn't=20 >know if the >failure is with this particular link or if it is further upstream. The >reason to remain in up/alloc is that it may see restored=20 >traffic over the >link if the failure is repaired upstream. > >Thanks, >Jonathan > >> -----Original Message----- >> From: George Young [mailto:george.young@edgeflow.com] >> Sent: Friday, September 07, 2001 10:13 AM >> To: ccamp@ops.ietf.org >> Subject: relationship of ChannelFail and Data Link FSM >>=20 >>=20 >> Regarding ID draft-ietf-ccamp-lmp-00.txt, I have a question about the >> ChannelFail message and the Data Link FSM.=20 >>=20 >> Firstly, my understanding is that when channel monitoring detects a >> failure, an LMP ChannelFail message is sent upstream, relative to the >> direction of transmission. A node which receives this message=20 >> and knows >> via monitoring that its upstream link is ok has localized=20 >the failure. >>=20 >> My question is does this cause any state change in the Data Link FSM, >> and if so, does it occur only in the node which localizes=20 >the failure, >> or all nodes which are aware of the failure. The current FSM does not >> seem to address this. >>=20 >> Regards, >> George. >> edgeflow Inc. >> 329 March Rd., Kanata, ON, Canada, K2K 2E1 >> phone: +1 613-270-9279 Ext 287 >> fax: +1 613-270-9268 >>=20 >>=20 > Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 10 Sep 2001 10:20:32 -0700 Message-ID: From: Jonathan Lang To: 'George Young' , ccamp@ops.ietf.org Subject: RE: relationship of ChannelFail and Data Link FSM Date: Mon, 10 Sep 2001 10:13:00 -0700 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" George, The upstream node should put the data link into the Down state. The downstream node remains in the Up/Alloc state since it doesn't know if the failure is with this particular link or if it is further upstream. The reason to remain in up/alloc is that it may see restored traffic over the link if the failure is repaired upstream. Thanks, Jonathan > -----Original Message----- > From: George Young [mailto:george.young@edgeflow.com] > Sent: Friday, September 07, 2001 10:13 AM > To: ccamp@ops.ietf.org > Subject: relationship of ChannelFail and Data Link FSM > > > Regarding ID draft-ietf-ccamp-lmp-00.txt, I have a question about the > ChannelFail message and the Data Link FSM. > > Firstly, my understanding is that when channel monitoring detects a > failure, an LMP ChannelFail message is sent upstream, relative to the > direction of transmission. A node which receives this message > and knows > via monitoring that its upstream link is ok has localized the failure. > > My question is does this cause any state change in the Data Link FSM, > and if so, does it occur only in the node which localizes the failure, > or all nodes which are aware of the failure. The current FSM does not > seem to address this. > > Regards, > George. > edgeflow Inc. > 329 March Rd., Kanata, ON, Canada, K2K 2E1 > phone: +1 613-270-9279 Ext 287 > fax: +1 613-270-9268 > > Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 07 Sep 2001 12:51:37 -0700 content-class: urn:content-classes:message Subject: relationship of RSVP-TE and LMP MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Date: Fri, 7 Sep 2001 15:47:28 -0400 Message-ID: Thread-Topic: relationship of RSVP-TE and LMP Thread-Index: AcE31ey3QHH4It+VQsGgnX9bKulvYw== From: "George Young" To: I have two questions about the relationship of RSVP-TE (draft-ietf-mpls-generalized-rsvp-te-03.txt) and LMP (draft-ietf-ccamp-lmp-00.txt). First of all, my understanding of the LMP ChannelActive message is that it is used to indicate to the downstream (relative to data flow direction) node that monitoring should be used. The scenario I have in mind is an RSVP call setup consisting of a bidirectional optical path. Presumably a switching node has to reserve data channels and associate them with the optical path when an RSVP PATH message arrives. 1) My first question is what is the timing of LMP ChannelActive messages relative to RSVP messages. Do ChannelActive messages flow downstream (relative to call establishment direction) more-or-less in phase with PATH messages, and upstream (relative to call establishment direction) more-or-less in phase with RESV messages? 2) My second question is what addtional information information do ChannelActive messages provide? If a link is to be monitored when associated with an active call, it appears the node already has this information. Regards, George. edgeflow Inc. 329 March Rd., Kanata, ON, Canada, K2K 2E1 phone: +1 613-270-9279 Ext 287 fax: +1 613-270-9268 Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 07 Sep 2001 10:17:36 -0700 content-class: urn:content-classes:message Subject: relationship of ChannelFail and Data Link FSM MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Date: Fri, 7 Sep 2001 13:12:32 -0400 Message-ID: Thread-Topic: relationship of ChannelFail and Data Link FSM Thread-Index: AcE3wEg4qCLXML1xRKalJCllRPwPtw== From: "George Young" To: Regarding ID draft-ietf-ccamp-lmp-00.txt, I have a question about the ChannelFail message and the Data Link FSM.=20 Firstly, my understanding is that when channel monitoring detects a failure, an LMP ChannelFail message is sent upstream, relative to the direction of transmission. A node which receives this message and knows via monitoring that its upstream link is ok has localized the failure. My question is does this cause any state change in the Data Link FSM, and if so, does it occur only in the node which localizes the failure, or all nodes which are aware of the failure. The current FSM does not seem to address this. Regards, George. edgeflow Inc. 329 March Rd., Kanata, ON, Canada, K2K 2E1 phone: +1 613-270-9279 Ext 287 fax: +1 613-270-9268 Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 06 Sep 2001 03:49:42 -0700 Message-Id: <200109061044.GAA19349@ietf.org> Mime-Version: 1.0 Content-Type: Multipart/Mixed; Boundary="NextPart" To: IETF-Announce: ; Cc: ccamp@ops.ietf.org From: Internet-Drafts@ietf.org Reply-to: Internet-Drafts@ietf.org Subject: I-D ACTION:draft-ietf-ccamp-gmpls-routing-00.txt Date: Thu, 06 Sep 2001 06:44:45 -0400 --NextPart A New Internet-Draft is available from the on-line Internet-Drafts directories. This draft is a work item of the Common Control and Measurement Plane Working Group of the IETF. Title : Routing Extensions in Support of Generalized MPLS Author(s) : K. Kompella et al. Filename : draft-ietf-ccamp-gmpls-routing-00.txt Pages : 20 Date : 05-Sep-01 This document specifies routing extensions in support of Generalized Multi-Protocol Label Switching (GMPLS). A URL for this Internet-Draft is: http://www.ietf.org/internet-drafts/draft-ietf-ccamp-gmpls-routing-00.txt To remove yourself from the IETF Announcement list, send a message to ietf-announce-request with the word unsubscribe in the body of the message. Internet-Drafts are also available by anonymous FTP. Login with the username "anonymous" and a password of your e-mail address. After logging in, type "cd internet-drafts" and then "get draft-ietf-ccamp-gmpls-routing-00.txt". A list of Internet-Drafts directories can be found in http://www.ietf.org/shadow.html or ftp://ftp.ietf.org/ietf/1shadow-sites.txt Internet-Drafts can also be obtained by e-mail. Send a message to: mailserv@ietf.org. In the body type: "FILE /internet-drafts/draft-ietf-ccamp-gmpls-routing-00.txt". NOTE: The mail server at ietf.org can return the document in MIME-encoded form by using the "mpack" utility. To use this feature, insert the command "ENCODING mime" before the "FILE" command. To decode the response(s), you will need "munpack" or a MIME-compliant mail reader. Different MIME-compliant mail readers exhibit different behavior, especially when dealing with "multipart" MIME messages (i.e. documents which have been split up into multiple messages), so check your local documentation on how to manipulate these messages. Below is the data which will enable a MIME compliant mail reader implementation to automatically retrieve the ASCII version of the Internet-Draft. --NextPart Content-Type: Multipart/Alternative; Boundary="OtherAccess" --OtherAccess Content-Type: Message/External-body; access-type="mail-server"; server="mailserv@ietf.org" Content-Type: text/plain Content-ID: <20010905125313.I-D@ietf.org> ENCODING mime FILE /internet-drafts/draft-ietf-ccamp-gmpls-routing-00.txt --OtherAccess Content-Type: Message/External-body; name="draft-ietf-ccamp-gmpls-routing-00.txt"; site="ftp.ietf.org"; access-type="anon-ftp"; directory="internet-drafts" Content-Type: text/plain Content-ID: <20010905125313.I-D@ietf.org> --OtherAccess-- --NextPart-- Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 06 Sep 2001 03:49:36 -0700 Message-Id: <200109061044.GAA19325@ietf.org> Mime-Version: 1.0 Content-Type: Multipart/Mixed; Boundary="NextPart" To: IETF-Announce: ; Cc: ccamp@ops.ietf.org From: Internet-Drafts@ietf.org Reply-to: Internet-Drafts@ietf.org Subject: I-D ACTION:draft-ietf-ccamp-ospf-gmpls-extensions-00.txt Date: Thu, 06 Sep 2001 06:44:40 -0400 --NextPart A New Internet-Draft is available from the on-line Internet-Drafts directories. This draft is a work item of the Common Control and Measurement Plane Working Group of the IETF. Title : OSPF Extensions in Support of Generalized MPLS Author(s) : K. Kompella et al. Filename : draft-ietf-ccamp-ospf-gmpls-extensions-00.txt Pages : 9 Date : 05-Sep-01 This document specifies encoding of extensions to the OSPF routing protocol in support of Generalized Multi-Protocol Label Switching (GMPLS). The description of the extensions is specified in [GMPLS- ROUTING]. A URL for this Internet-Draft is: http://www.ietf.org/internet-drafts/draft-ietf-ccamp-ospf-gmpls-extensions-00.txt To remove yourself from the IETF Announcement list, send a message to ietf-announce-request with the word unsubscribe in the body of the message. Internet-Drafts are also available by anonymous FTP. Login with the username "anonymous" and a password of your e-mail address. After logging in, type "cd internet-drafts" and then "get draft-ietf-ccamp-ospf-gmpls-extensions-00.txt". A list of Internet-Drafts directories can be found in http://www.ietf.org/shadow.html or ftp://ftp.ietf.org/ietf/1shadow-sites.txt Internet-Drafts can also be obtained by e-mail. Send a message to: mailserv@ietf.org. In the body type: "FILE /internet-drafts/draft-ietf-ccamp-ospf-gmpls-extensions-00.txt". NOTE: The mail server at ietf.org can return the document in MIME-encoded form by using the "mpack" utility. To use this feature, insert the command "ENCODING mime" before the "FILE" command. To decode the response(s), you will need "munpack" or a MIME-compliant mail reader. Different MIME-compliant mail readers exhibit different behavior, especially when dealing with "multipart" MIME messages (i.e. documents which have been split up into multiple messages), so check your local documentation on how to manipulate these messages. Below is the data which will enable a MIME compliant mail reader implementation to automatically retrieve the ASCII version of the Internet-Draft. --NextPart Content-Type: Multipart/Alternative; Boundary="OtherAccess" --OtherAccess Content-Type: Message/External-body; access-type="mail-server"; server="mailserv@ietf.org" Content-Type: text/plain Content-ID: <20010905125303.I-D@ietf.org> ENCODING mime FILE /internet-drafts/draft-ietf-ccamp-ospf-gmpls-extensions-00.txt --OtherAccess Content-Type: Message/External-body; name="draft-ietf-ccamp-ospf-gmpls-extensions-00.txt"; site="ftp.ietf.org"; access-type="anon-ftp"; directory="internet-drafts" Content-Type: text/plain Content-ID: <20010905125303.I-D@ietf.org> --OtherAccess-- --NextPart-- Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 05 Sep 2001 09:08:20 -0700 Message-ID: <3B964C7C.595E40F6@research.att.com> Date: Wed, 05 Sep 2001 12:02:04 -0400 From: Guangzhi Li MIME-Version: 1.0 To: Hang Liu Cc: Fong Liaw , Eric Gray , John Drake , ccamp@ops.ietf.org, mpls@UU.NET, gli@research.att.com Subject: Re: Question on GMPLS contentions resolution Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hang: Yah. I see the point. As long as this clarification is addressed, the rule (2) may not be needed since setup priority has to be greater than the holding priority to preempt an established LSP. One question: assume LSP A with Path message from node 1 and node 2 and LSP B with Resv message from node 2 to node 1. Node 1 has cross-conneced for LSP A and node 2 has cross-connected for LSP B. The setup prority is equal to the holding priority for both LSP A and LSP B. It is easy to resolve the contention on node 2. How about node 1? LSP A has been setup on node 1 but LSP B has not. Resv message is required to use the setup priority to preempt other established LSPs. What is the harm to put a rule without confliction with previous definitions. Otherwise, other clarifications are required. Regards, Guangzhi Hang Liu wrote: > I think that Fong's points are valid. We can assume that "RESV" message > means that the LSP has been set up (at least within the downstream node). > There are the holding priority associated with it. The PATH message for the > other LSP can win the contentinon only if its set-up priority > the holding > priority of another LSP. > > Hang > > -----Original Message----- > From: Guangzhi Li [mailto:gli@research.att.com] > Sent: Wednesday, September 05, 2001 9:22 AM > To: Fong Liaw > Cc: Eric Gray; John Drake; ccamp@ops.ietf.org; mpls@UU.NET > Subject: Re: Question on GMPLS contentions resolution > > Fong: > > Would you please explain more why (2) is not needed if (1) is changed ? > IF (holding priority == setup priority), how to resolve it ? > > By the way, how to interperate holding priority and setup priority should be > explained when they are defined. > > (3) is an extension of current GMPLS specification. Current GMPLS clearly > says that > this rule deals with two bi-directional LSPs. > > Guangzhi > > Fong Liaw wrote: > > > Guangzhi > > > > GMPLS includes SESSION_ATTRIBUTE object, > > but the current text on contention resolution > > does not take it into account. So I think adding > > (1) as clarification is needed, however, I would > > change (1) to > > > > "(1) Based on the LSPs' hold and setup priority." > > > > I might remove (2) if it is understood that > > receiving a Resv means hold priority should be > > used in (1). > > > > We already have (3), so that is fine. > > > > -Fong > > > > --- Guangzhi Li wrote: > > > Hi, all: > > > > > > I posted the contention resolution question 10 days > > > ago. After some fruitful > > > discussions with John Drake, Eric Gray, Fong Liaw, > > > Hang Liu, I propose the > > > following three rules to resolve any contention in > > > GMPLS although there is no > > > signaling correctness issue. The following rules > > > should applied in sequence to > > > both bi-directional and uni-directional LSPs: > > > > > > (1) Higher priority LSP wins lower priority LSP. > > > (2) LSP with Reply (Resv in RSVP) message wins LSP > > > with Request > > > (Path in RSVP) message. > > > (3) LSP with the higher node ID assigning the label > > > wins the contention > > > if race conditions happen between two LSPs with > > > Request messages. > > > > > > Since race conditions won't happen between two LSPs > > > with Reply messages, > > > the above three rules solve all possible contention > > > scenarios. > > > > > > Proposed modification of the GMPLS functional > > > specification, text from > > > section 4.2: > > > > > > --------------------------------------------------------------------------- > > > > > > By allowing bi-directional LSP setup and suggested > > > label in the GMPLS > > > framework, contention for labels may occur between > > > two bidirectional > > > LSP setup requests, or between an uni-directional > > > and a bi-directional > > > LSP, traveling in opposite directions. This > > > contention occurs when both sides > > > allocate the same resources (labels) at effectively > > > the same time. If there is no > > > restriction on the labels that can be used for > > > bidirectional LSPs and if > > > there are alternate resources, downstream label > > > assignment will apply > > > and there is no contention. However, if there is a > > > restriction on the > > > labels that can be used for the bidirectional LSPs > > > (for example, if they > > > must be physically coupled on a single I/O card), or > > > if there are no > > > more resources available, then the contention must > > > be resolved by other > > > means. Although contention resolution is not a > > > mandatory requirement for > > > signaling correctness, contention resolution is > > > strongly recommended > > > for performance optimization. To resolve contention, > > > the following three rules > > > SHOULD be applied in sequence: > > > (1) A higher priority LSP wins over a lower priority > > > LSP. > > > (2) LSP with a Reply message wins over a LSP with a > > > Request > > > message. > > > (3) LSP with higher node ID assigning the label wins > > > the contention if > > > race conditions happen between two LSPs with Request > > > messages. > > > ........ > > > > > > The rest of the text from this section should be > > > updated accordingly. > > > > > > Thanks, > > > > > > Guangzhi > > > > > > > > > > Dear GMPLS authors and all experts: > > > > > > > > During GMPLS last call, I posted the same question > > > on the mailing list > > > > without response. Please somebody spend a little > > > time and check with the > > > > following example? Something seems not clear when > > > the current GMPLS > > > > contention resution schemes are applied on a > > > framework with both > > > > bi-directional and uni-directional LSPs. Your > > > clarification is very much > > > > appreciated. > > > > > > > > Sincerely, > > > > > > > > Guangzhi > > > > > > > > > > > > > > ---------------------------------------------------------------------------- > --- > > > > > > > > The issue arises because contention is resolved > > > between bi-directional > > > > LSPs by the node with the higher node index while > > > for uni-directional > > > > LSPs, contention is resolved by the downstream > > > node. Consider the > > > > following example of two nodes with paired, > > > bi-directional interfaces > > > > (i.e., a transmitter/receiver pair of ports). > > > Node 1 with ID=100 and > > > > node 2 with ID = 50. Node 1 uses label 1 for the > > > transmitter port and > > > > label 2 for the receiver port; node 2 uses label 4 > > > for the transmitter > > > > port and label 3 for the receiver port. We > > > assume that a > > > > bi-directional LSP requires a single I/O > > > interface. > > > > > > > > We consider two LSPs - a uni-directional LSP (LSP > > > A) and a > > > > bi-directional LSP (LSP B). Both LSPs are going > > > from node 1 > > > > to node 2, with the uni-directional LSP setup > > > request arriving marginally before > > > > > > > > the bi-directional LSP. LSP A does not use a > > > suggested label, and thus > > > > is assigned a label (port) by node 2. Label 3 is > > > assigned, > > > > corresponding to label 1 at node 1. At the same > > > time, for LSP B (the > > > > bi-directional LSP) node 1 assigns label 2, with > > > suggested label 1. > > > > Because node 1 has the higher node ID, node 2 will > > > assume (due to the > > > > contention resolution rule for bi-directional > > > LSPs) that LSP B wins the > > > > contention and thus label 3 is assigned to LSP B. > > > Thus label 1 at node > > > > 1 (label 3 at node 2) has been assigned to two > > > different LSPs. Both > > > > LSPs have "won" the contention. > > > > > > > > > > __________________________________________________ > > Do You Yahoo!? > > Get email alerts & NEW webcam video instant messaging with Yahoo! > Messenger > > http://im.yahoo.com Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 05 Sep 2001 08:20:30 -0700 Message-ID: <3B963F31.649677D3@sandburst.com> Date: Wed, 05 Sep 2001 11:05:21 -0400 From: Eric Gray MIME-Version: 1.0 To: Guangzhi Li Cc: Fong Liaw , John Drake , ccamp@ops.ietf.org, mpls@UU.NET Subject: Re: Question on GMPLS contentions resolution Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Guangzhi, The reference to setup and holding priorities would require a reference to [RSVP-TE] and [CR-LDP] references already included in the draft. Whther or not these need to be specifically called out again is a call for the authors - assuming that they include any of your proposed changes. -- Eric Gray You wrote: > Fong: > > Would you please explain more why (2) is not needed if (1) is changed ? > IF (holding priority == setup priority), how to resolve it ? > > By the way, how to interperate holding priority and setup priority should be > explained when they are defined. > > (3) is an extension of current GMPLS specification. Current GMPLS clearly says that > this rule deals with two bi-directional LSPs. > > Guangzhi > > Fong Liaw wrote: > > > Guangzhi > > > > GMPLS includes SESSION_ATTRIBUTE object, > > but the current text on contention resolution > > does not take it into account. So I think adding > > (1) as clarification is needed, however, I would > > change (1) to > > > > "(1) Based on the LSPs' hold and setup priority." > > > > I might remove (2) if it is understood that > > receiving a Resv means hold priority should be > > used in (1). > > > > We already have (3), so that is fine. > > > > -Fong > > > > --- Guangzhi Li wrote: > > > Hi, all: > > > > > > I posted the contention resolution question 10 days > > > ago. After some fruitful > > > discussions with John Drake, Eric Gray, Fong Liaw, > > > Hang Liu, I propose the > > > following three rules to resolve any contention in > > > GMPLS although there is no > > > signaling correctness issue. The following rules > > > should applied in sequence to > > > both bi-directional and uni-directional LSPs: > > > > > > (1) Higher priority LSP wins lower priority LSP. > > > (2) LSP with Reply (Resv in RSVP) message wins LSP > > > with Request > > > (Path in RSVP) message. > > > (3) LSP with the higher node ID assigning the label > > > wins the contention > > > if race conditions happen between two LSPs with > > > Request messages. > > > > > > Since race conditions won't happen between two LSPs > > > with Reply messages, > > > the above three rules solve all possible contention > > > scenarios. > > > > > > Proposed modification of the GMPLS functional > > > specification, text from > > > section 4.2: > > > > > --------------------------------------------------------------------------- > > > > > > By allowing bi-directional LSP setup and suggested > > > label in the GMPLS > > > framework, contention for labels may occur between > > > two bidirectional > > > LSP setup requests, or between an uni-directional > > > and a bi-directional > > > LSP, traveling in opposite directions. This > > > contention occurs when both sides > > > allocate the same resources (labels) at effectively > > > the same time. If there is no > > > restriction on the labels that can be used for > > > bidirectional LSPs and if > > > there are alternate resources, downstream label > > > assignment will apply > > > and there is no contention. However, if there is a > > > restriction on the > > > labels that can be used for the bidirectional LSPs > > > (for example, if they > > > must be physically coupled on a single I/O card), or > > > if there are no > > > more resources available, then the contention must > > > be resolved by other > > > means. Although contention resolution is not a > > > mandatory requirement for > > > signaling correctness, contention resolution is > > > strongly recommended > > > for performance optimization. To resolve contention, > > > the following three rules > > > SHOULD be applied in sequence: > > > (1) A higher priority LSP wins over a lower priority > > > LSP. > > > (2) LSP with a Reply message wins over a LSP with a > > > Request > > > message. > > > (3) LSP with higher node ID assigning the label wins > > > the contention if > > > race conditions happen between two LSPs with Request > > > messages. > > > ........ > > > > > > The rest of the text from this section should be > > > updated accordingly. > > > > > > Thanks, > > > > > > Guangzhi > > > > > > > > > > Dear GMPLS authors and all experts: > > > > > > > > During GMPLS last call, I posted the same question > > > on the mailing list > > > > without response. Please somebody spend a little > > > time and check with the > > > > following example? Something seems not clear when > > > the current GMPLS > > > > contention resution schemes are applied on a > > > framework with both > > > > bi-directional and uni-directional LSPs. Your > > > clarification is very much > > > > appreciated. > > > > > > > > Sincerely, > > > > > > > > Guangzhi > > > > > > > > > > > > > ------------------------------------------------------------------------------- > > > > > > > > The issue arises because contention is resolved > > > between bi-directional > > > > LSPs by the node with the higher node index while > > > for uni-directional > > > > LSPs, contention is resolved by the downstream > > > node. Consider the > > > > following example of two nodes with paired, > > > bi-directional interfaces > > > > (i.e., a transmitter/receiver pair of ports). > > > Node 1 with ID=100 and > > > > node 2 with ID = 50. Node 1 uses label 1 for the > > > transmitter port and > > > > label 2 for the receiver port; node 2 uses label 4 > > > for the transmitter > > > > port and label 3 for the receiver port. We > > > assume that a > > > > bi-directional LSP requires a single I/O > > > interface. > > > > > > > > We consider two LSPs - a uni-directional LSP (LSP > > > A) and a > > > > bi-directional LSP (LSP B). Both LSPs are going > > > from node 1 > > > > to node 2, with the uni-directional LSP setup > > > request arriving marginally before > > > > > > > > the bi-directional LSP. LSP A does not use a > > > suggested label, and thus > > > > is assigned a label (port) by node 2. Label 3 is > > > assigned, > > > > corresponding to label 1 at node 1. At the same > > > time, for LSP B (the > > > > bi-directional LSP) node 1 assigns label 2, with > > > suggested label 1. > > > > Because node 1 has the higher node ID, node 2 will > > > assume (due to the > > > > contention resolution rule for bi-directional > > > LSPs) that LSP B wins the > > > > contention and thus label 3 is assigned to LSP B. > > > Thus label 1 at node > > > > 1 (label 3 at node 2) has been assigned to two > > > different LSPs. Both > > > > LSPs have "won" the contention. > > > > > > > > > > __________________________________________________ > > Do You Yahoo!? > > Get email alerts & NEW webcam video instant messaging with Yahoo! Messenger > > http://im.yahoo.com Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 05 Sep 2001 08:18:11 -0700 Message-ID: <3B963E7C.3D31E615@sandburst.com> Date: Wed, 05 Sep 2001 11:02:20 -0400 From: Eric Gray MIME-Version: 1.0 To: Guangzhi Li Cc: Fong Liaw , John Drake , ccamp@ops.ietf.org, mpls@UU.NET Subject: Re: Question on GMPLS contentions resolution Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Guangzhi, In any case where there is a tie between holding priority and setup priority, holding priority wins. Setup priority MUST be higher (i.e. - have a lower value) than holding priority for setup to win. This is not clearly stated anywhere in the current RSVP-TE draft, nor in the RFC (2751) that it refers to, nor in the CR-LDP draft. However, it may be directly inferred from the fact that the setup priority is allowed to be equal to the hold/holding priority for the same LSP being setup but may not be greater. The stated reason for this is to preclude an LSP from being preempted by another LSP having identical priorities and then preempting that LSP in turn. Clearly this intent would not be satisfied if a tie resulted in preemption. -- Eric Gray wrote: > Fong: > > Would you please explain more why (2) is not needed if (1) is changed ? > IF (holding priority == setup priority), how to resolve it ? > > By the way, how to interperate holding priority and setup priority should be > explained when they are defined. > > (3) is an extension of current GMPLS specification. Current GMPLS clearly says that > this rule deals with two bi-directional LSPs. > > Guangzhi > > Fong Liaw wrote: > > > Guangzhi > > > > GMPLS includes SESSION_ATTRIBUTE object, > > but the current text on contention resolution > > does not take it into account. So I think adding > > (1) as clarification is needed, however, I would > > change (1) to > > > > "(1) Based on the LSPs' hold and setup priority." > > > > I might remove (2) if it is understood that > > receiving a Resv means hold priority should be > > used in (1). > > > > We already have (3), so that is fine. > > > > -Fong > > > > --- Guangzhi Li wrote: > > > Hi, all: > > > > > > I posted the contention resolution question 10 days > > > ago. After some fruitful > > > discussions with John Drake, Eric Gray, Fong Liaw, > > > Hang Liu, I propose the > > > following three rules to resolve any contention in > > > GMPLS although there is no > > > signaling correctness issue. The following rules > > > should applied in sequence to > > > both bi-directional and uni-directional LSPs: > > > > > > (1) Higher priority LSP wins lower priority LSP. > > > (2) LSP with Reply (Resv in RSVP) message wins LSP > > > with Request > > > (Path in RSVP) message. > > > (3) LSP with the higher node ID assigning the label > > > wins the contention > > > if race conditions happen between two LSPs with > > > Request messages. > > > > > > Since race conditions won't happen between two LSPs > > > with Reply messages, > > > the above three rules solve all possible contention > > > scenarios. > > > > > > Proposed modification of the GMPLS functional > > > specification, text from > > > section 4.2: > > > > > --------------------------------------------------------------------------- > > > > > > By allowing bi-directional LSP setup and suggested > > > label in the GMPLS > > > framework, contention for labels may occur between > > > two bidirectional > > > LSP setup requests, or between an uni-directional > > > and a bi-directional > > > LSP, traveling in opposite directions. This > > > contention occurs when both sides > > > allocate the same resources (labels) at effectively > > > the same time. If there is no > > > restriction on the labels that can be used for > > > bidirectional LSPs and if > > > there are alternate resources, downstream label > > > assignment will apply > > > and there is no contention. However, if there is a > > > restriction on the > > > labels that can be used for the bidirectional LSPs > > > (for example, if they > > > must be physically coupled on a single I/O card), or > > > if there are no > > > more resources available, then the contention must > > > be resolved by other > > > means. Although contention resolution is not a > > > mandatory requirement for > > > signaling correctness, contention resolution is > > > strongly recommended > > > for performance optimization. To resolve contention, > > > the following three rules > > > SHOULD be applied in sequence: > > > (1) A higher priority LSP wins over a lower priority > > > LSP. > > > (2) LSP with a Reply message wins over a LSP with a > > > Request > > > message. > > > (3) LSP with higher node ID assigning the label wins > > > the contention if > > > race conditions happen between two LSPs with Request > > > messages. > > > ........ > > > > > > The rest of the text from this section should be > > > updated accordingly. > > > > > > Thanks, > > > > > > Guangzhi > > > > > > > > > > Dear GMPLS authors and all experts: > > > > > > > > During GMPLS last call, I posted the same question > > > on the mailing list > > > > without response. Please somebody spend a little > > > time and check with the > > > > following example? Something seems not clear when > > > the current GMPLS > > > > contention resution schemes are applied on a > > > framework with both > > > > bi-directional and uni-directional LSPs. Your > > > clarification is very much > > > > appreciated. > > > > > > > > Sincerely, > > > > > > > > Guangzhi > > > > > > > > > > > > > ------------------------------------------------------------------------------- > > > > > > > > The issue arises because contention is resolved > > > between bi-directional > > > > LSPs by the node with the higher node index while > > > for uni-directional > > > > LSPs, contention is resolved by the downstream > > > node. Consider the > > > > following example of two nodes with paired, > > > bi-directional interfaces > > > > (i.e., a transmitter/receiver pair of ports). > > > Node 1 with ID=100 and > > > > node 2 with ID = 50. Node 1 uses label 1 for the > > > transmitter port and > > > > label 2 for the receiver port; node 2 uses label 4 > > > for the transmitter > > > > port and label 3 for the receiver port. We > > > assume that a > > > > bi-directional LSP requires a single I/O > > > interface. > > > > > > > > We consider two LSPs - a uni-directional LSP (LSP > > > A) and a > > > > bi-directional LSP (LSP B). Both LSPs are going > > > from node 1 > > > > to node 2, with the uni-directional LSP setup > > > request arriving marginally before > > > > > > > > the bi-directional LSP. LSP A does not use a > > > suggested label, and thus > > > > is assigned a label (port) by node 2. Label 3 is > > > assigned, > > > > corresponding to label 1 at node 1. At the same > > > time, for LSP B (the > > > > bi-directional LSP) node 1 assigns label 2, with > > > suggested label 1. > > > > Because node 1 has the higher node ID, node 2 will > > > assume (due to the > > > > contention resolution rule for bi-directional > > > LSPs) that LSP B wins the > > > > contention and thus label 3 is assigned to LSP B. > > > Thus label 1 at node > > > > 1 (label 3 at node 2) has been assigned to two > > > different LSPs. Both > > > > LSPs have "won" the contention. > > > > > > > > > > __________________________________________________ > > Do You Yahoo!? > > Get email alerts & NEW webcam video instant messaging with Yahoo! Messenger > > http://im.yahoo.com Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 05 Sep 2001 08:00:12 -0700 Message-ID: <0CD7D73C734BD411AEE100B0D022040402088957@mail1.tellium.com> From: Hang Liu To: "'Guangzhi Li'" , Fong Liaw Cc: Eric Gray , John Drake , ccamp@ops.ietf.org, mpls@UU.NET Subject: RE: Question on GMPLS contentions resolution Date: Wed, 5 Sep 2001 10:45:22 -0400 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" I think that Fong's points are valid. We can assume that "RESV" message means that the LSP has been set up (at least within the downstream node). There are the holding priority associated with it. The PATH message for the other LSP can win the contentinon only if its set-up priority > the holding priority of another LSP. Hang -----Original Message----- From: Guangzhi Li [mailto:gli@research.att.com] Sent: Wednesday, September 05, 2001 9:22 AM To: Fong Liaw Cc: Eric Gray; John Drake; ccamp@ops.ietf.org; mpls@UU.NET Subject: Re: Question on GMPLS contentions resolution Fong: Would you please explain more why (2) is not needed if (1) is changed ? IF (holding priority == setup priority), how to resolve it ? By the way, how to interperate holding priority and setup priority should be explained when they are defined. (3) is an extension of current GMPLS specification. Current GMPLS clearly says that this rule deals with two bi-directional LSPs. Guangzhi Fong Liaw wrote: > Guangzhi > > GMPLS includes SESSION_ATTRIBUTE object, > but the current text on contention resolution > does not take it into account. So I think adding > (1) as clarification is needed, however, I would > change (1) to > > "(1) Based on the LSPs' hold and setup priority." > > I might remove (2) if it is understood that > receiving a Resv means hold priority should be > used in (1). > > We already have (3), so that is fine. > > -Fong > > --- Guangzhi Li wrote: > > Hi, all: > > > > I posted the contention resolution question 10 days > > ago. After some fruitful > > discussions with John Drake, Eric Gray, Fong Liaw, > > Hang Liu, I propose the > > following three rules to resolve any contention in > > GMPLS although there is no > > signaling correctness issue. The following rules > > should applied in sequence to > > both bi-directional and uni-directional LSPs: > > > > (1) Higher priority LSP wins lower priority LSP. > > (2) LSP with Reply (Resv in RSVP) message wins LSP > > with Request > > (Path in RSVP) message. > > (3) LSP with the higher node ID assigning the label > > wins the contention > > if race conditions happen between two LSPs with > > Request messages. > > > > Since race conditions won't happen between two LSPs > > with Reply messages, > > the above three rules solve all possible contention > > scenarios. > > > > Proposed modification of the GMPLS functional > > specification, text from > > section 4.2: > > > --------------------------------------------------------------------------- > > > > By allowing bi-directional LSP setup and suggested > > label in the GMPLS > > framework, contention for labels may occur between > > two bidirectional > > LSP setup requests, or between an uni-directional > > and a bi-directional > > LSP, traveling in opposite directions. This > > contention occurs when both sides > > allocate the same resources (labels) at effectively > > the same time. If there is no > > restriction on the labels that can be used for > > bidirectional LSPs and if > > there are alternate resources, downstream label > > assignment will apply > > and there is no contention. However, if there is a > > restriction on the > > labels that can be used for the bidirectional LSPs > > (for example, if they > > must be physically coupled on a single I/O card), or > > if there are no > > more resources available, then the contention must > > be resolved by other > > means. Although contention resolution is not a > > mandatory requirement for > > signaling correctness, contention resolution is > > strongly recommended > > for performance optimization. To resolve contention, > > the following three rules > > SHOULD be applied in sequence: > > (1) A higher priority LSP wins over a lower priority > > LSP. > > (2) LSP with a Reply message wins over a LSP with a > > Request > > message. > > (3) LSP with higher node ID assigning the label wins > > the contention if > > race conditions happen between two LSPs with Request > > messages. > > ........ > > > > The rest of the text from this section should be > > updated accordingly. > > > > Thanks, > > > > Guangzhi > > > > > > > Dear GMPLS authors and all experts: > > > > > > During GMPLS last call, I posted the same question > > on the mailing list > > > without response. Please somebody spend a little > > time and check with the > > > following example? Something seems not clear when > > the current GMPLS > > > contention resution schemes are applied on a > > framework with both > > > bi-directional and uni-directional LSPs. Your > > clarification is very much > > > appreciated. > > > > > > Sincerely, > > > > > > Guangzhi > > > > > > > > > ---------------------------------------------------------------------------- --- > > > > > > The issue arises because contention is resolved > > between bi-directional > > > LSPs by the node with the higher node index while > > for uni-directional > > > LSPs, contention is resolved by the downstream > > node. Consider the > > > following example of two nodes with paired, > > bi-directional interfaces > > > (i.e., a transmitter/receiver pair of ports). > > Node 1 with ID=100 and > > > node 2 with ID = 50. Node 1 uses label 1 for the > > transmitter port and > > > label 2 for the receiver port; node 2 uses label 4 > > for the transmitter > > > port and label 3 for the receiver port. We > > assume that a > > > bi-directional LSP requires a single I/O > > interface. > > > > > > We consider two LSPs - a uni-directional LSP (LSP > > A) and a > > > bi-directional LSP (LSP B). Both LSPs are going > > from node 1 > > > to node 2, with the uni-directional LSP setup > > request arriving marginally before > > > > > > the bi-directional LSP. LSP A does not use a > > suggested label, and thus > > > is assigned a label (port) by node 2. Label 3 is > > assigned, > > > corresponding to label 1 at node 1. At the same > > time, for LSP B (the > > > bi-directional LSP) node 1 assigns label 2, with > > suggested label 1. > > > Because node 1 has the higher node ID, node 2 will > > assume (due to the > > > contention resolution rule for bi-directional > > LSPs) that LSP B wins the > > > contention and thus label 3 is assigned to LSP B. > > Thus label 1 at node > > > 1 (label 3 at node 2) has been assigned to two > > different LSPs. Both > > > LSPs have "won" the contention. > > > > > > __________________________________________________ > Do You Yahoo!? > Get email alerts & NEW webcam video instant messaging with Yahoo! Messenger > http://im.yahoo.com Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 05 Sep 2001 06:23:12 -0700 Message-ID: <3B9626E7.DACEEF2E@research.att.com> Date: Wed, 05 Sep 2001 09:21:43 -0400 From: Guangzhi Li MIME-Version: 1.0 To: Fong Liaw Cc: Eric Gray , John Drake , ccamp@ops.ietf.org, mpls@UU.NET Subject: Re: Question on GMPLS contentions resolution Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Fong: Would you please explain more why (2) is not needed if (1) is changed ? IF (holding priority == setup priority), how to resolve it ? By the way, how to interperate holding priority and setup priority should be explained when they are defined. (3) is an extension of current GMPLS specification. Current GMPLS clearly says that this rule deals with two bi-directional LSPs. Guangzhi Fong Liaw wrote: > Guangzhi > > GMPLS includes SESSION_ATTRIBUTE object, > but the current text on contention resolution > does not take it into account. So I think adding > (1) as clarification is needed, however, I would > change (1) to > > "(1) Based on the LSPs' hold and setup priority." > > I might remove (2) if it is understood that > receiving a Resv means hold priority should be > used in (1). > > We already have (3), so that is fine. > > -Fong > > --- Guangzhi Li wrote: > > Hi, all: > > > > I posted the contention resolution question 10 days > > ago. After some fruitful > > discussions with John Drake, Eric Gray, Fong Liaw, > > Hang Liu, I propose the > > following three rules to resolve any contention in > > GMPLS although there is no > > signaling correctness issue. The following rules > > should applied in sequence to > > both bi-directional and uni-directional LSPs: > > > > (1) Higher priority LSP wins lower priority LSP. > > (2) LSP with Reply (Resv in RSVP) message wins LSP > > with Request > > (Path in RSVP) message. > > (3) LSP with the higher node ID assigning the label > > wins the contention > > if race conditions happen between two LSPs with > > Request messages. > > > > Since race conditions won't happen between two LSPs > > with Reply messages, > > the above three rules solve all possible contention > > scenarios. > > > > Proposed modification of the GMPLS functional > > specification, text from > > section 4.2: > > > --------------------------------------------------------------------------- > > > > By allowing bi-directional LSP setup and suggested > > label in the GMPLS > > framework, contention for labels may occur between > > two bidirectional > > LSP setup requests, or between an uni-directional > > and a bi-directional > > LSP, traveling in opposite directions. This > > contention occurs when both sides > > allocate the same resources (labels) at effectively > > the same time. If there is no > > restriction on the labels that can be used for > > bidirectional LSPs and if > > there are alternate resources, downstream label > > assignment will apply > > and there is no contention. However, if there is a > > restriction on the > > labels that can be used for the bidirectional LSPs > > (for example, if they > > must be physically coupled on a single I/O card), or > > if there are no > > more resources available, then the contention must > > be resolved by other > > means. Although contention resolution is not a > > mandatory requirement for > > signaling correctness, contention resolution is > > strongly recommended > > for performance optimization. To resolve contention, > > the following three rules > > SHOULD be applied in sequence: > > (1) A higher priority LSP wins over a lower priority > > LSP. > > (2) LSP with a Reply message wins over a LSP with a > > Request > > message. > > (3) LSP with higher node ID assigning the label wins > > the contention if > > race conditions happen between two LSPs with Request > > messages. > > ........ > > > > The rest of the text from this section should be > > updated accordingly. > > > > Thanks, > > > > Guangzhi > > > > > > > Dear GMPLS authors and all experts: > > > > > > During GMPLS last call, I posted the same question > > on the mailing list > > > without response. Please somebody spend a little > > time and check with the > > > following example? Something seems not clear when > > the current GMPLS > > > contention resution schemes are applied on a > > framework with both > > > bi-directional and uni-directional LSPs. Your > > clarification is very much > > > appreciated. > > > > > > Sincerely, > > > > > > Guangzhi > > > > > > > > > ------------------------------------------------------------------------------- > > > > > > The issue arises because contention is resolved > > between bi-directional > > > LSPs by the node with the higher node index while > > for uni-directional > > > LSPs, contention is resolved by the downstream > > node. Consider the > > > following example of two nodes with paired, > > bi-directional interfaces > > > (i.e., a transmitter/receiver pair of ports). > > Node 1 with ID=100 and > > > node 2 with ID = 50. Node 1 uses label 1 for the > > transmitter port and > > > label 2 for the receiver port; node 2 uses label 4 > > for the transmitter > > > port and label 3 for the receiver port. We > > assume that a > > > bi-directional LSP requires a single I/O > > interface. > > > > > > We consider two LSPs - a uni-directional LSP (LSP > > A) and a > > > bi-directional LSP (LSP B). Both LSPs are going > > from node 1 > > > to node 2, with the uni-directional LSP setup > > request arriving marginally before > > > > > > the bi-directional LSP. LSP A does not use a > > suggested label, and thus > > > is assigned a label (port) by node 2. Label 3 is > > assigned, > > > corresponding to label 1 at node 1. At the same > > time, for LSP B (the > > > bi-directional LSP) node 1 assigns label 2, with > > suggested label 1. > > > Because node 1 has the higher node ID, node 2 will > > assume (due to the > > > contention resolution rule for bi-directional > > LSPs) that LSP B wins the > > > contention and thus label 3 is assigned to LSP B. > > Thus label 1 at node > > > 1 (label 3 at node 2) has been assigned to two > > different LSPs. Both > > > LSPs have "won" the contention. > > > > > > __________________________________________________ > Do You Yahoo!? > Get email alerts & NEW webcam video instant messaging with Yahoo! Messenger > http://im.yahoo.com Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 04 Sep 2001 21:28:48 -0700 Message-ID: <20010905042619.53626.qmail@web14701.mail.yahoo.com> Date: Tue, 4 Sep 2001 21:26:19 -0700 (PDT) From: Fong Liaw Subject: Re: Question on GMPLS contentions resolution To: Guangzhi Li , Eric Gray Cc: John Drake , ccamp@ops.ietf.org, mpls@UU.NET MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Guangzhi GMPLS includes SESSION_ATTRIBUTE object, but the current text on contention resolution does not take it into account. So I think adding (1) as clarification is needed, however, I would change (1) to "(1) Based on the LSPs' hold and setup priority." I might remove (2) if it is understood that receiving a Resv means hold priority should be used in (1). We already have (3), so that is fine. -Fong --- Guangzhi Li wrote: > Hi, all: > > I posted the contention resolution question 10 days > ago. After some fruitful > discussions with John Drake, Eric Gray, Fong Liaw, > Hang Liu, I propose the > following three rules to resolve any contention in > GMPLS although there is no > signaling correctness issue. The following rules > should applied in sequence to > both bi-directional and uni-directional LSPs: > > (1) Higher priority LSP wins lower priority LSP. > (2) LSP with Reply (Resv in RSVP) message wins LSP > with Request > (Path in RSVP) message. > (3) LSP with the higher node ID assigning the label > wins the contention > if race conditions happen between two LSPs with > Request messages. > > Since race conditions won't happen between two LSPs > with Reply messages, > the above three rules solve all possible contention > scenarios. > > Proposed modification of the GMPLS functional > specification, text from > section 4.2: > --------------------------------------------------------------------------- > > By allowing bi-directional LSP setup and suggested > label in the GMPLS > framework, contention for labels may occur between > two bidirectional > LSP setup requests, or between an uni-directional > and a bi-directional > LSP, traveling in opposite directions. This > contention occurs when both sides > allocate the same resources (labels) at effectively > the same time. If there is no > restriction on the labels that can be used for > bidirectional LSPs and if > there are alternate resources, downstream label > assignment will apply > and there is no contention. However, if there is a > restriction on the > labels that can be used for the bidirectional LSPs > (for example, if they > must be physically coupled on a single I/O card), or > if there are no > more resources available, then the contention must > be resolved by other > means. Although contention resolution is not a > mandatory requirement for > signaling correctness, contention resolution is > strongly recommended > for performance optimization. To resolve contention, > the following three rules > SHOULD be applied in sequence: > (1) A higher priority LSP wins over a lower priority > LSP. > (2) LSP with a Reply message wins over a LSP with a > Request > message. > (3) LSP with higher node ID assigning the label wins > the contention if > race conditions happen between two LSPs with Request > messages. > ........ > > The rest of the text from this section should be > updated accordingly. > > Thanks, > > Guangzhi > > > > Dear GMPLS authors and all experts: > > > > During GMPLS last call, I posted the same question > on the mailing list > > without response. Please somebody spend a little > time and check with the > > following example? Something seems not clear when > the current GMPLS > > contention resution schemes are applied on a > framework with both > > bi-directional and uni-directional LSPs. Your > clarification is very much > > appreciated. > > > > Sincerely, > > > > Guangzhi > > > > > ------------------------------------------------------------------------------- > > > > The issue arises because contention is resolved > between bi-directional > > LSPs by the node with the higher node index while > for uni-directional > > LSPs, contention is resolved by the downstream > node. Consider the > > following example of two nodes with paired, > bi-directional interfaces > > (i.e., a transmitter/receiver pair of ports). > Node 1 with ID=100 and > > node 2 with ID = 50. Node 1 uses label 1 for the > transmitter port and > > label 2 for the receiver port; node 2 uses label 4 > for the transmitter > > port and label 3 for the receiver port. We > assume that a > > bi-directional LSP requires a single I/O > interface. > > > > We consider two LSPs - a uni-directional LSP (LSP > A) and a > > bi-directional LSP (LSP B). Both LSPs are going > from node 1 > > to node 2, with the uni-directional LSP setup > request arriving marginally before > > > > the bi-directional LSP. LSP A does not use a > suggested label, and thus > > is assigned a label (port) by node 2. Label 3 is > assigned, > > corresponding to label 1 at node 1. At the same > time, for LSP B (the > > bi-directional LSP) node 1 assigns label 2, with > suggested label 1. > > Because node 1 has the higher node ID, node 2 will > assume (due to the > > contention resolution rule for bi-directional > LSPs) that LSP B wins the > > contention and thus label 3 is assigned to LSP B. > Thus label 1 at node > > 1 (label 3 at node 2) has been assigned to two > different LSPs. Both > > LSPs have "won" the contention. > > __________________________________________________ Do You Yahoo!? Get email alerts & NEW webcam video instant messaging with Yahoo! Messenger http://im.yahoo.com Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 04 Sep 2001 18:53:05 -0700 Message-ID: <319146423A05D411B66800D0B73E9419240129@EXCH_BVTN> From: "Anantha, Vic" To: ccamp@ops.ietf.org Subject: RE: Question on GMPLS contentions resolution Date: Tue, 4 Sep 2001 15:47:59 -0700 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Could somebody point me to where I can read up more about the node IDs that are being mentioned here. Are they IP addresses that are interpreted as 32 bit unsigned integers? Vik -----Original Message----- From: Guangzhi Li [mailto:gli@research.att.com] Sent: Monday, August 27, 2001 2:13 PM To: Fong Liaw Cc: ccamp@ops.ietf.org; mpls@UU.NET Subject: Re: Question on GMPLS contentions resolution Fong: Uni-directional LSP uses RESV to assign labels. IF a GMPLS network runs both uni-directional and bi-directional together and uses the same port pool, then a PATH and a RESV may cross each other. I think that it should be legal based on current GMPLS specification. IF this scenario happens, GMPLS has to provide one standard way to handle this situation. Your suggestion works but it still involves the implementations of node 1 and node 2. Both nodes should interperate the same way. That means RESV message always wins PATH message. No problem. I assume that the I/O ports of bi-didrectional LSP are paired together, then suggested_label = upstream_label and suggested_label is hard binding. Note that contention happens in either this case or lack of resource. If suggested_label could be different upstream_label, no contention problem at all. -- Guangzhi Fong Liaw wrote: > Guangzhi > > The GMPLS contention resolution applies to the scenario > that two Path messages cross each other. Your example > have Resv and Path cross each other so the rule in the > contention resolution does not directly apply and is more > of a policy and implementation issue. > > Here is one possible solution and is within the > interpretation of the draft, IMHO. here it is. > If node 2 already sends Resv to node 1, LSP A should > be considered established and therefore LSP B should > be rejected by node 2. If node 2 hasn't sent back Resv, > LSP A is still in establishing mode and node 2 can resolve > the contention internally by rejecting LSP A or LSP B, based > on their setting up priority or it can change LSP A's > label and reprogram the cross-connect without ever noticed > by any external node. In both cases, node 2 is the one > that resolve this contention. Hope this helps. > > p.s. I think you meant to use LABEL_SET since > Suggested label is a suggestion, not a hard binding. > > Regards > -Fong > > > -----Original Message----- > > From: Guangzhi Li [mailto:gli@research.att.com] > > Sent: Monday, August 27, 2001 6:17 AM > > To: ccamp@ops.ietf.org; mpls@UU.NET > > Cc: gli@research.att.com > > Subject: Question on GMPLS contentions resolution > > > > > > Dear GMPLS authors and all experts: > > > > During GMPLS last call, I posted the same question on the > > mailing list > > without response. Please somebody spend a little time and > > check with the > > following example? Something seems not clear when the current GMPLS > > contention resution schemes are applied on a framework with both > > bi-directional and uni-directional LSPs. Your clarification > > is very much > > appreciated. > > > > Sincerely, > > > > Guangzhi > > > > ------------------------------------------------------------- > > ------------------------------------- > > > > The issue arises because contention is resolved between > > bi-directional > > LSPs by the node with the higher node index while for uni-directional > > LSPs, contention is resolved by the downstream node. Consider the > > following example of two nodes with paired, bi-directional interfaces > > (i.e., a transmitter/receiver pair of ports). Node 1 with ID=100 and > > node 2 with ID = 50. Node 1 uses label 1 for the > > transmitter port and > > label 2 for the receiver port; node 2 uses label 4 for the > > transmitter > > port and label 3 for the receiver port. We assume that a > > bi-directional LSP requires a single I/O interface. > > > > We consider two LSPs - a uni-directional LSP (LSP A) and a > > bi-directional LSP (LSP B). Both LSPs are going from node 1 > > to node 2, > > with the uni-directional LSP setup request arriving marginally before > > the bi-directional LSP. LSP A does not use a suggested > > label, and thus > > is assigned a label (port) by node 2. Label 3 is assigned, > > corresponding to label 1 at node 1. At the same time, for LSP B (the > > bi-directional LSP) node 1 assigns label 2, with suggested label 1. > > Because node 1 has the higher node ID, node 2 will assume (due to the > > contention resolution rule for bi-directional LSPs) that LSP > > B wins the > > contention and thus label 3 is assigned to LSP B. Thus > > label 1 at node > > 1 (label 3 at node 2) has been assigned to two different LSPs. Both > > LSPs have "won" the contention. > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 04 Sep 2001 12:50:30 -0700 Message-ID: <001201c13585$74e5e330$8e452c42@tpidl6z5ynrc41> From: "S Kalipatnapu" To: "Bart Batsleer WX22 54174" , Subject: Re: Questions on OSPF extensions for TE Date: Tue, 4 Sep 2001 17:06:21 -0400 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Bart, Section 2.5.5 on Traffic Engineering Metric, states the following: > >The Traffic Engineering Metric sub-TLV specifies the link metric for > traffic engineering purposes. This metric may be different than the > standard OSPF link metric. > The "may be" in the last sentence may be indicating what you are looking for. Regards, --Sitaram ----- Original Message ----- From: "Bart Batsleer WX22 54174" To: Sent: Tuesday, September 04, 2001 11:27 AM Subject: Questions on OSPF extensions for TE > Hello, > > I have 2 questions related to the draft-katz-yeung-ospf-traffic-05 > > 1) In case the traffic engineering metric sub-tlv is not present in the > link > tlv, does the traffic engineering SPF calculation have to rely on the > standard OSPF link metric ? > > 2) The behaviour is stated in the corresponding isis draft > (draft-ietf-isis-traffic-02), is there a > reason why this is not stated in the OSPF draft ? > > Thanks in advance, > Bart > > > Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 04 Sep 2001 11:59:34 -0700 Message-Id: <200109041856.f84Iuf011384@merlot.juniper.net> To: "Atkinson, Stephen" cc: ccamp@ops.ietf.org Subject: Re: draft-ietf-mpls-generalized-signaling-05 Interface Identification MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-ID: <66481.999629801.1@juniper.net> Date: Tue, 04 Sep 2001 11:56:41 -0700 From: Yakov Rekhter Steve, > draft-ietf-mpls-generalized-signaling-05 section 9.1 defines a new > Interface_ID TLV which > allows for specifying Types of: > > IPv4 > IPv6 > IF_INDEX > COMPONENT_IF_DOWNSTREAM > COMPONENT_IF_UPSTREAM > > Why the inclusion of Components? > Shouldn't components be signaled as per draft-kompella-mpls-bundle-05? > In the presence of bundling is it valid to have an Interface_ID TLV which > specifies the components? Wouldn't it be redundant? > > Having two ways of signaling the same information may simply create problems > with interoperability. Agreed wrt "two ways". And that is why we just submitted a revised version of the bundling Internet Draft that uses just IF_ID. Yakov. Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 04 Sep 2001 08:29:26 -0700 Message-ID: <3B94F2EE.F8C406EA@alcatel.be> Date: Tue, 04 Sep 2001 17:27:42 +0200 From: Bart Batsleer WX22 54174 Organization: Alcatel Telecom - Antwerp - Belgium MIME-Version: 1.0 To: ccamp@ops.ietf.org Subject: Questions on OSPF extensions for TE Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hello, I have 2 questions related to the draft-katz-yeung-ospf-traffic-05 1) In case the traffic engineering metric sub-tlv is not present in the link tlv, does the traffic engineering SPF calculation have to rely on the standard OSPF link metric ? 2) The behaviour is stated in the corresponding isis draft (draft-ietf-isis-traffic-02), is there a reason why this is not stated in the OSPF draft ? Thanks in advance, Bart Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 04 Sep 2001 07:47:56 -0700 Message-ID: <3B94EA88.708CB4C@research.att.com> Date: Tue, 04 Sep 2001 10:51:52 -0400 From: Guangzhi Li MIME-Version: 1.0 To: Eric Gray Cc: John Drake , Fong Liaw , ccamp@ops.ietf.org, mpls@UU.NET Subject: Re: Question on GMPLS contentions resolution Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi, all: I posted the contention resolution question 10 days ago. After some fruitful discussions with John Drake, Eric Gray, Fong Liaw, Hang Liu, I propose the following three rules to resolve any contention in GMPLS although there is no signaling correctness issue. The following rules should applied in sequence to both bi-directional and uni-directional LSPs: (1) Higher priority LSP wins lower priority LSP. (2) LSP with Reply (Resv in RSVP) message wins LSP with Request (Path in RSVP) message. (3) LSP with the higher node ID assigning the label wins the contention if race conditions happen between two LSPs with Request messages. Since race conditions won't happen between two LSPs with Reply messages, the above three rules solve all possible contention scenarios. Proposed modification of the GMPLS functional specification, text from section 4.2: --------------------------------------------------------------------------- By allowing bi-directional LSP setup and suggested label in the GMPLS framework, contention for labels may occur between two bidirectional LSP setup requests, or between an uni-directional and a bi-directional LSP, traveling in opposite directions. This contention occurs when both sides allocate the same resources (labels) at effectively the same time. If there is no restriction on the labels that can be used for bidirectional LSPs and if there are alternate resources, downstream label assignment will apply and there is no contention. However, if there is a restriction on the labels that can be used for the bidirectional LSPs (for example, if they must be physically coupled on a single I/O card), or if there are no more resources available, then the contention must be resolved by other means. Although contention resolution is not a mandatory requirement for signaling correctness, contention resolution is strongly recommended for performance optimization. To resolve contention, the following three rules SHOULD be applied in sequence: (1) A higher priority LSP wins over a lower priority LSP. (2) LSP with a Reply message wins over a LSP with a Request message. (3) LSP with higher node ID assigning the label wins the contention if race conditions happen between two LSPs with Request messages. ........ The rest of the text from this section should be updated accordingly. Thanks, Guangzhi > Dear GMPLS authors and all experts: > > During GMPLS last call, I posted the same question on the mailing list > without response. Please somebody spend a little time and check with the > following example? Something seems not clear when the current GMPLS > contention resution schemes are applied on a framework with both > bi-directional and uni-directional LSPs. Your clarification is very much > appreciated. > > Sincerely, > > Guangzhi > > ------------------------------------------------------------------------------- > > The issue arises because contention is resolved between bi-directional > LSPs by the node with the higher node index while for uni-directional > LSPs, contention is resolved by the downstream node. Consider the > following example of two nodes with paired, bi-directional interfaces > (i.e., a transmitter/receiver pair of ports). Node 1 with ID=100 and > node 2 with ID = 50. Node 1 uses label 1 for the transmitter port and > label 2 for the receiver port; node 2 uses label 4 for the transmitter > port and label 3 for the receiver port. We assume that a > bi-directional LSP requires a single I/O interface. > > We consider two LSPs - a uni-directional LSP (LSP A) and a > bi-directional LSP (LSP B). Both LSPs are going from node 1 > to node 2, with the uni-directional LSP setup request arriving marginally before > > the bi-directional LSP. LSP A does not use a suggested label, and thus > is assigned a label (port) by node 2. Label 3 is assigned, > corresponding to label 1 at node 1. At the same time, for LSP B (the > bi-directional LSP) node 1 assigns label 2, with suggested label 1. > Because node 1 has the higher node ID, node 2 will assume (due to the > contention resolution rule for bi-directional LSPs) that LSP B wins the > contention and thus label 3 is assigned to LSP B. Thus label 1 at node > 1 (label 3 at node 2) has been assigned to two different LSPs. Both > LSPs have "won" the contention. Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 04 Sep 2001 05:50:18 -0700 Message-ID: <87009604743AD411B1F600508BA0F959632448@XOVER.dedham.mindspeed.com> From: "Atkinson, Stephen" To: ccamp@ops.ietf.org Subject: draft-ietf-mpls-generalized-signaling-05 Interface Identification Date: Tue, 4 Sep 2001 08:48:44 -0400 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" draft-ietf-mpls-generalized-signaling-05 section 9.1 defines a new Interface_ID TLV which allows for specifying Types of: IPv4 IPv6 IF_INDEX COMPONENT_IF_DOWNSTREAM COMPONENT_IF_UPSTREAM Why the inclusion of Components? Shouldn't components be signaled as per draft-kompella-mpls-bundle-05? In the presence of bundling is it valid to have an Interface_ID TLV which specifies the components? Wouldn't it be redundant? Having two ways of signaling the same information may simply create problems with interoperability. Thanks, steve atkinson Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 04 Sep 2001 05:50:12 -0700 Message-ID: <87009604743AD411B1F600508BA0F959632446@XOVER.dedham.mindspeed.com> From: "Atkinson, Stephen" To: "'ccamp@ops.ietf.org'" Subject: Generalized MPLS, Bundling, and bi-directional LSPs with uni-dire ctional interfaces Date: Tue, 4 Sep 2001 08:46:12 -0400 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" To quote from draft-ietf-mpls-generalized-rsvp-te-04 section 8.1, "In all cases but bundling [MPLS-BUNDLE] the upstream interface is implied by the downstream interface. For bundling, the path sender explicitly identifies the component interface used in each direction.". I'm missing something here. This means, in the absence of Bundling, interfaces must always be bi-directional if used for setting up bi-directional LSPs. Why is Bundling different in this regard? Is there something specific to bundling requiring the ability to setup bi-directional LSPs on uni-directional interfaces, which isn't required in the absence of bundling? For example Node A and Node B are connected by 4 fibers. Two fibers on Node A are outgoing and two are incoming. In both cases below we're setting up a bi-directional LSP. Case 1: The fibers are bundled as per MPLS-BUNDLE. In this case the component to use in each direction is specified by DOWNSTREAM_COMPONENT_INTERFACE_ID and UPSTREAM_COMPONENT_INTERFACE_ID as per MPLS-BUNDLE. Case 2: The fibers are not bundled. In this case we can use the Interface_ID TLV as per section 9.1 of draft-ietf-mpls-generalized-signaling-05 (Type = 1, 2 or 3) to specify the downstream fiber to use. There is no mechanism for specifying the upstream fiber. Case 2 won't work unless there's an apriori coupling of an incoming and outgoing fiber. If this is the case, why is this coupling not used in the presence of bundling? Thanks, steve atkinson Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 04 Sep 2001 05:50:07 -0700 Message-ID: <87009604743AD411B1F600508BA0F959632447@XOVER.dedham.mindspeed.com> From: "Atkinson, Stephen" To: ccamp@ops.ietf.org Subject: Questions on draft-kompella-mpls-bundle-05 Date: Tue, 4 Sep 2001 08:47:28 -0400 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" I've been reading the bundle-05 spec and have a couple questions. Bundling (section 5.2) requires that the upstream node specify the component interface used in each direction when setting up bi-directional LSPs. But only the "Explicit Indication by Component Interface ID" mechanism of signaling specifies how this is accomplished. How can you specify both components using "Explicit Indication by IP Address"? Are the IP source and destination addresses expected to be the means of specifying both the Upstream and Downstream components? How can you specify both components using "Implicit Indication"? I don't see a way to make this work. Is it intended that Explicit Indication by IP Address and Implicit Indication only setup bi-directional LSPs using bi-directional links? Thanks, steve atkinson