From sip-security-admin@ietf.org  Mon Sep  3 14:50:43 2001
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA13255
	for <sip-security-archive@odin.ietf.org>; Mon, 3 Sep 2001 14:50:43 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA26536;
	Mon, 3 Sep 2001 14:51:25 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA26496
	for <sip-security@ns.ietf.org>; Mon, 3 Sep 2001 14:51:23 -0400 (EDT)
Received: from fep01-app.kolumbus.fi (fep01-0.kolumbus.fi [193.229.0.41])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA13246;
	Mon, 3 Sep 2001 14:49:56 -0400 (EDT)
Received: from jariws1 ([212.54.16.198]) by fep01-app.kolumbus.fi
          (InterMail vM.5.01.03.08 201-253-122-118-108-20010628) with SMTP
          id <20010903185114.HQCQ11276.fep01-app.kolumbus.fi@jariws1>;
          Mon, 3 Sep 2001 21:51:14 +0300
Message-ID: <007701c134a9$6dbe7400$8a1b6e0a@arenanet.fi>
From: "Jari Arkko" <jari.arkko@kolumbus.fi>
To: "Sriram Parameswar" <sriramp@nortelnetworks.com>, <sip@ietf.org>
Cc: <jari.arkko@ericsson.fi>, <valerius@nortelnetworks.com>,
        <tiina.s.koskinen@nokia.com>, <sip-security@ietf.org>
References: <9A9367D1556AD21182C40000F80930AB04411D78@crchy28b.us.nortel.com>
Date: Mon, 3 Sep 2001 21:51:23 +0300
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 5.50.4133.2400
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4133.2400
Content-Transfer-Encoding: 7bit
Subject: [Sip-security] Re: [Sip] SIP Negotiate
Sender: sip-security-admin@ietf.org
Errors-To: sip-security-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Security Issues for the SIP protocol <sip-security.ietf.org>
X-BeenThere: sip-security@ietf.org
Content-Transfer-Encoding: 7bit

RE: [Sipping] SIP Negotiate
I agree that there is a need for this kind of functionality. However,
when you say in the beginning "Examples of parameters that
... include ... message integrity mechanisms, encryption
algorithms, etc.", I'd like to note that this what we need in
terms of security parameters is a _secure_ negotiation of them.
As you say yourself in section 10.1 of
http://www.obsidian97.com/draft-spbs-sip-negotiate-00.txt :

> One of the intended uses for the negotiation mechanism is to set
> up authentication and message integrity algorithms. Thus the 
> NEGOTIATE method itself is not subject to these protections - this
> makes the recipient vulnerable to the Man-in-the-middle attacks. 
> Views are solicited to mitigate this security consideration.

This is a crucial point to handle correctly. Otherwise the peers
may be fooled onto using insecure algorithms or perhaps even
no algorithms at all. It is also interesting to note that the answer
can't really be "integrity protect the NEGOTIATE" because
that would mean an integrity protection algorithm would have
to be applied before one is selected!

We've done some work on this issue earlier in the context of
SIP usage in 3G networks, and came up with some requirements.
These included the following:

* MUST be able to choose different integrity algorithms,
  MAY be able to choose other security parameters,
  encryption algorithms, and security protocols.
* It MUST be possible to protect the parameter negotiation
  against attackers who do not have the secret authentication
  keys (and can't forge hashes in real-time).
* MUST not add new roundtrips (i.e. done in parallel with
  other setup such as registration or authentication). In particular,
  it would not be nice if we had separate negotiation phases for security,
  compression, sip itself, and only then we'd be able to do
  something.

Traditionally, security protocols have solved secure negotiation
of parameters by hashing/signing the offers with authentication
keys or keys derived from authentication. For instance, IKE
negotiates parameters in cleartext but in a later phase a keyed
hash is taken from the original parameter offers and responses,
and sent to the other side as a proof that nothing was changed in
the packet's path.

Something similar needs to happen here as well. We've had
the idea that once integrity protection can be turned on, you'll
repeat the offers under it, and only after this you can start to
trust and use what was negotiated. (Integrity protection can be
turned on basically after something has given you keys
and you have agreed on the selected algorithm.)

Jari Arkko
Ericsson




_______________________________________________
Sip-security mailing list
Sip-security@ietf.org
http://www1.ietf.org/mailman/listinfo/sip-security


From sip-security-admin@ietf.org  Fri Sep  7 07:35:40 2001
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA07091
	for <sip-security-archive@odin.ietf.org>; Fri, 7 Sep 2001 07:35:40 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA06688;
	Fri, 7 Sep 2001 07:36:23 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA06655
	for <sip-security@optimus.ietf.org>; Fri, 7 Sep 2001 07:36:22 -0400 (EDT)
Received: from gorilla.mchh.siemens.de (gorilla.mchh.siemens.de [194.138.158.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA07075
	for <sip-security@ietf.org>; Fri, 7 Sep 2001 07:34:54 -0400 (EDT)
Received: from moody.mchh.siemens.de (mail2.mchh.siemens.de [194.138.158.226])
	by gorilla.mchh.siemens.de (8.9.3/8.9.3) with ESMTP id NAA21704;
	Fri, 7 Sep 2001 13:36:17 +0200 (MET DST)
Received: from mchh273e.demchh201e.icn.siemens.de ([139.21.200.83])
	by moody.mchh.siemens.de (8.9.1/8.9.1) with ESMTP id NAA06212;
	Fri, 7 Sep 2001 13:36:19 +0200 (MET DST)
Received: by MCHH273E with Internet Mail Service (5.5.2653.19)
	id <SMDGV5V8>; Fri, 7 Sep 2001 13:35:38 +0200
Message-ID: <5316E8083EC4D411957F0008C71EEE34A71D49@MCHH227E>
From: Euchner Martin <Martin.Euchner@icn.siemens.de>
To: "'Michael Thomas'" <mat@cisco.com>,
        Jari Arkko
	 <Jari.Arkko@lmf.ericsson.se>,
        Euchner Martin
	 <Martin.Euchner@icn.siemens.de>
Cc: sip-security@ietf.org
Subject: AW: [Sip-security] Re: Proxy routed requirements in draft-thomas-
	sip-sec-framework-00.txt
Date: Fri, 7 Sep 2001 13:35:34 +0200 
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="ISO-8859-1"
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id HAA06656
Sender: sip-security-admin@ietf.org
Errors-To: sip-security-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Security Issues for the SIP protocol <sip-security.ietf.org>
X-BeenThere: sip-security@ietf.org
Content-Transfer-Encoding: 8bit

I would definitely agree to say that non-repudiation is a non issue for SIP for quite some time. This could be marked as for further study or so, if it would ever be an issue for SIP.

Denial-of-Service threats are more interesting. It is certainly true, that there are SIP DoS attacks which we should bother. As an example, Jonathan mentioned the DoS through amplification problem. Such direct SIP DoS attacks are something the SIP protocol should counter. I'm not certain, as to whether such DoS attacks have already been analyzed thoroughly and comprehensively for SIP; probably some more investigations are necessary.

However, there are also various other indirect DoS attacks, which may cause troubles at the SIP layer. Many of them would actually target the underlying network, some would be massive integrity or masquerade attacks, spamming, message flooding.

It is quite difficult to find a sound answer as to whether
-	such attacks should be detected or even prevented by non-SIP means; e.g. in the network, within routers, by firewalls, by network monitors, by ISPs etc. This should be the primary, first-line-of defense against DoS Attacks.
-	or SIP can at least do something simple to detect those DoS attacks that cannot be thwarted by above means. This would serve as a second-line of defense and complement to above one.
-	And where to spend the countermeasures.

	My personal opinion is, that SIP
-	should make sure that the protocol design addresses DoS threats: protocol state transitions should be done in a safe way, message amplification does not occur and such things. This is even an area that does not require cryptographic security.
-	some simple, low computational effort countermeasures: cookies perhaps as in IKE, symmetric integrity check values in SIP messages or other adequate cryptography. It might even be sufficient, if a weak DoS countermeasure be available. This would fit the 2nd line-of-defense strategy.


	The crucial points are:
-	What are the most serious DoS attacks for SIP? Which ones are just annoying and which one would really cause serious harm (financial fraud)?
-	How much are DoS countermeasures allowed to cost? What are the acceptable constraints such as additional protocol overhead, roundtrips just for DoS detection/prevention, that wouldn't be required if DoS attacks were a non-issue?
-	Which DoS attacks can we counter within some acceptable constraints and what are those constraints?
-	Which other DoS attacks would be considered out-of-scope for some reason and must be tackled otherwise?

Finally, SIP DoS prevention could be an issue in certain environments and scenarios, while in other ones, it would be a minor or even no issue at all. This would mean, that SIP DoS should be a separate, optional on/off feature.

In any way, DoS threats are an issue, but perhaps not the most urgent one for SIP security.

Kind Regards

Martin Euchner.
-----------------------------------------------------------------------
| Dipl.-Inf.                     Phone: +49 89 722 55790
| Martin Euchner                 Fax  : +49 89 722 46841
| Siemens AG
| ICN M SR 3                     mailto:Martin.Euchner@icn.siemens.de <mailto:Martin.Euchner@icn.siemens.de> 
|                                mailto:martin.euchner@ties.itu.int <mailto:martin.euchner@ties.itu.int> 
| Hofmannstr. 51                 Intranet: http://intranet.icn.siemens.de/marketing/sr/pages/122/122_euchner.htm <http://intranet.icn.siemens.de/marketing/cs27/> 
| D-81359 Muenchen               Internet: http://www.siemens.de <http://www.siemens.de> 
| __________________
| Germany     
-----------------------------------------------------------------------

	-----Ursprüngliche Nachricht-----
	Von:	Michael Thomas [SMTP:mat@cisco.com]
	Gesendet am:	Freitag, 31. August 2001 18:23
	An:	Jari Arkko
	Cc:	mat@cisco.com; sip-security@ietf.org
	Betreff:	[Sip-security] Re: Proxy routed requirements in draft-thomas-sip-sec-framework-00.txt

	Jari Arkko writes:
	 > 
	 > Some further discussion: comparing Michael's
	 > draft and draft-ietf-sip-security-requirements-00.txt,
	 > I see that the former doesn't handle availability
	 > and non-repudiation requirements at all. Yet these
	 > are mentioned in the latter. I believe availability
	 > and DoS is a concern, and should perhaps be treated
	 > somewhere. Non-repudiation on the other hand is
	 > propably a can of worms we don't want to open...

	   Yes, I really don't want to go down the 
	   non-repudiation path. Maybe we just need to
	   say that it is an explicit non-goal. DoS is
	   a more fertile ground, but as we all know is
	   an open ended problem. In order to include
	   discussion of DoS requirements, I think we
	   first need to create some artificials bounds
	   on the discussion so that we have some chance
	   of creating a set of requirements which 
	   a potential solution can objectively measure
	   whether it meets. Ie, open ended statements like
	   "should be DoS resistant" aren't helpful.

	   How we frame this is an interesting question.
	   For outside attackers, we clearly need go no
	   farther than countering a pure flooding attack.
	   From the inside, however, is a lot trickier, and
	   involves potential tradeoffs (tradeoffs that we
	   probably don't want to make at this point, I'll
	   add). 

	   Maybe if we rounded up a bunch of potential DoS
	   attacks we could see a pattern so that we could
	   extract what our expectations are for SIP's 
	   resilience to them is?

	 > Also, in 3.2.5 you talk about the case for
	 > non-adjacent proxies, and wonder if there is
	 > any need for security measures in this
	 > interface. My comment on this is that
	 > protocol-wise you wouldn't even know if
	 > someone who e.g. sent some protected bits
	 > to you was a proxy or an end-node. So,
	 > if we provide tools to get security on the
	 > user->intermediate interface, then non-adjacent
	 > proxies can take advantage of the same
	 > mechanisms. (This comment doesn't of course
	 > say anything about the requirements for such use.
	 > It is perhaps not unthinkable to have a
	 > proxy that would handle security for you
	 > in certain situations.)

	   If there are no requirements, or the requirements
	   are identical, that's fine -- good even, I'd say.
	   The reason I created the fully connected graph
	   was to assess whether that was indeed the case.
	   It appears to me to be the case, but I'd like 
	   to see if other people have sceanrios I hadn't
	   considered, or don't fit that model well.

			Mike

	_______________________________________________
	Sip-security mailing list
	Sip-security@ietf.org
	http://www1.ietf.org/mailman/listinfo/sip-security

_______________________________________________
Sip-security mailing list
Sip-security@ietf.org
http://www1.ietf.org/mailman/listinfo/sip-security


From sip-security-admin@ietf.org  Fri Sep  7 11:06:09 2001
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA16653
	for <sip-security-archive@odin.ietf.org>; Fri, 7 Sep 2001 11:06:09 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA13673;
	Fri, 7 Sep 2001 11:06:50 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA13638
	for <sip-security@optimus.ietf.org>; Fri, 7 Sep 2001 11:06:49 -0400 (EDT)
Received: from mail1.dynamicsoft.com ([63.113.40.10])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA16642
	for <sip-security@ietf.org>; Fri, 7 Sep 2001 11:05:21 -0400 (EDT)
Received: from DYN-EXCH-001.dynamicsoft.com (dyn-exch-001 [63.113.44.7])
	by mail1.dynamicsoft.com (8.12.0.Beta7/8.12.0.Beta7) with ESMTP id f87F5ICj027318;
	Fri, 7 Sep 2001 11:05:21 -0400 (EDT)
Received: by DYN-EXCH-001.dynamicsoft.com with Internet Mail Service (5.5.2653.19)
	id <R9F8YWA0>; Fri, 7 Sep 2001 11:06:11 -0400
Message-ID: <B65B4F8437968F488A01A940B21982BF020D67E4@DYN-EXCH-001.dynamicsoft.com>
From: Jonathan Rosenberg <jdrosen@dynamicsoft.com>
To: "'Euchner Martin'" <Martin.Euchner@icn.siemens.de>,
        "'Michael Thomas'"
	 <mat@cisco.com>,
        Jari Arkko <Jari.Arkko@lmf.ericsson.se>
Cc: sip-security@ietf.org
Subject: RE: [Sip-security] Re: Proxy routed requirements in draft-thomas-
	 sip-sec-framework-00.txt
Date: Fri, 7 Sep 2001 11:06:04 -0400 
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: sip-security-admin@ietf.org
Errors-To: sip-security-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Security Issues for the SIP protocol <sip-security.ietf.org>
X-BeenThere: sip-security@ietf.org



 

> -----Original Message-----
> From: Euchner Martin [mailto:Martin.Euchner@icn.siemens.de]
> Sent: Friday, September 07, 2001 7:36 AM
> To: 'Michael Thomas'; Jari Arkko; Euchner Martin
> Cc: sip-security@ietf.org
> Subject: AW: [Sip-security] Re: Proxy routed requirements in
> draft-thomas- sip-sec-framework-00.txt
> 
> Denial-of-Service threats are more interesting. It is 
> certainly true, that there are SIP DoS attacks which we 
> should bother. As an example, Jonathan mentioned the DoS 
> through amplification problem. Such direct SIP DoS attacks 
> are something the SIP protocol should counter. I'm not 
> certain, as to whether such DoS attacks have already been 
> analyzed thoroughly and comprehensively for SIP; probably 
> some more investigations are necessary.

I am sure more investigation is needed. We already know of a few problems.
One is that you simply send an INVITE without credentials to a server, and
it challenges you with 401. Based on the SIP spec rules, it will resend this
401, and therefore maintain state to do so. This can be used as an
amplification attack, and as a DoS attack. The counter, which is not well
documented, is to run the authentication "statelessly" - i.e., the server
sends the 401/407, but does not retransmit. This works OK for INVITE so long
as the server doesn't send a 1xx first. 

> It is quite difficult to find a sound answer as to whether
> -	such attacks should be detected or even prevented by 
> non-SIP means; e.g. in the network, within routers, by 
> firewalls, by network monitors, by ISPs etc. This should be 
> the primary, first-line-of defense against DoS Attacks.

Source address filtering helps with alot of these.

> 	My personal opinion is, that SIP
> -	should make sure that the protocol design addresses DoS 
> threats: protocol state transitions should be done in a safe 
> way, message amplification does not occur and such things. 
> This is even an area that does not require cryptographic security.

Right. We do have a real problem though right now. The stateless 401/407
appraoch above helps with proxies that have some kind of subscriber
relationships so that they can authenticate them. Many proxies, however,
cannot realistically authenticate the originator of a request (inbound
proxies, for example). So, these are subject to serious DoS attacks. One way
to resolve that is to always use something like TLS or TCP even, but this is
expensive in terms of connection management. We need to have a simple,
STATELESS, deployable, workable solution for a proxy to challenge the
previous hop, for the purposes of identifying that it is actually listening
on the address it advertised. 

I have a proposal on how to accomplish this. It works if we make the right
decision on request merging (i.e., merged requests are each treated as
independent transactions). Here is the proposal.

A server (uas, proxy, whatever) receives a request. It statelessly sends a
new response, 4xx, with a new challenge header:

Hop-Authenticate: null, realm=foobar.com, nonce=193347fakasd0asd8


THis uses the null authentication algorithm that Henning has proposed. The
previous hop receives this, and ACKs (the ACK is ignored by the proxy since
it generated a stateless response). Then, it resubmits the request (as if it
forked, much like as if the 4xx were a redirect instead), but now includes a
hop authorization header that merely reflects the nonce:

Hop-Authorization: null,realm=foodbar.com,nonce=193347fakasd0asd8

The server gets this, verifies the nonce is one it recently generated, and
then processes the request.

Since the resubmitted request has the same r-uri, request merging needs to
properly handle this case (i.e, not reject the request) which is why  I
mentioned this above.

I believe this approach works through stateless proxies, which is very good.


There is a backwards compatibility issue that is serious. Existing proxies
won't know to resubmit the request. We can come up with a way for a new
proxy to indicate that it does or doesn't support this extension in a
request. But, that introduces the attack back in! For the dos prevention to
work, a proxy would always need to be challenging the requests it gets.

There is also a huge penalty in this approach in terms of call setup delays,
potentially tripling them. In that regard, tls is better since that penalty
is bigger, but just once, as opposed to for each call setup. 

I know I am guilty of jumping to implementations before requirements work,
but we know that dos attack prevention is a requirement, and this is a big
problem.

I'll note that you could also use hop-authentication to allow a UAS to
actually challenge its inbound proxy using basic or digest... might be
useful to solve a few of the cases in Mike's draft. Combine that with James'
improved digest mechanism, and you have a reasonably lightweight way to do
proxy to proxy, and proxy to UAS authentication and message integrity.

> 
> 	The crucial points are:
> -	What are the most serious DoS attacks for SIP? Which 
> ones are just annoying and which one would really cause 
> serious harm (financial fraud)?

I think the most serious one is causing establishment of state for
unauthenticated transactions, which is the problem I have addressed above.

> In any way, DoS threats are an issue, but perhaps not the 
> most urgent one for SIP security.

I disagree. I think DoS threats are one of the most important security
problems. 

-Jonathan R.

---
Jonathan D. Rosenberg, Ph.D.                72 Eagle Rock Ave.
Chief Scientist                             First Floor
dynamicsoft                                 East Hanover, NJ 07936
jdrosen@dynamicsoft.com                     FAX:   (973) 952-5050
http://www.jdrosen.net                      PHONE: (973) 952-5000
http://www.dynamicsoft.com

_______________________________________________
Sip-security mailing list
Sip-security@ietf.org
http://www1.ietf.org/mailman/listinfo/sip-security


From sip-security-admin@ietf.org  Sat Sep  8 01:05:31 2001
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id BAA07897
	for <sip-security-archive@odin.ietf.org>; Sat, 8 Sep 2001 01:05:30 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id BAA11616;
	Sat, 8 Sep 2001 01:06:22 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id BAA11554
	for <sip-security@optimus.ietf.org>; Sat, 8 Sep 2001 01:06:19 -0400 (EDT)
Received: from fep01-app.kolumbus.fi (fep01-0.kolumbus.fi [193.229.0.41])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id BAA07848;
	Sat, 8 Sep 2001 01:04:52 -0400 (EDT)
Received: from jariws1 ([212.54.16.229]) by fep01-app.kolumbus.fi
          (InterMail vM.5.01.03.08 201-253-122-118-108-20010628) with SMTP
          id <20010908050617.XDND11276.fep01-app.kolumbus.fi@jariws1>;
          Sat, 8 Sep 2001 08:06:17 +0300
Message-ID: <015701c13824$01d503c0$8a1b6e0a@arenanet.fi>
From: "Jari Arkko" <jari.arkko@kolumbus.fi>
To: "Sriram Parameswar" <sriramp@nortelnetworks.com>,
        "'Jonathan Rosenberg'" <jdrosen@dynamicsoft.com>,
        "Sean Olson" <sean.olson@ericsson.com>, <sipping@ietf.org>,
        "Brian Stucker" <bstucker@nortelnetworks.com>
Cc: <sip-security@ietf.org>
References: <9A9367D1556AD21182C40000F80930AB04411DA1@crchy28b.us.nortel.com>
Date: Sat, 8 Sep 2001 08:06:24 +0300
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 5.50.4133.2400
X-MIMEOLE: Produced By Microsoft MimeOLE V5.50.4133.2400
Content-Transfer-Encoding: 7bit
Subject: [Sip-security] Re: [Sipping] SIP Negotiate
Sender: sip-security-admin@ietf.org
Errors-To: sip-security-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Security Issues for the SIP protocol <sip-security.ietf.org>
X-BeenThere: sip-security@ietf.org
Content-Transfer-Encoding: 7bit

RE: [Sipping] SIP Negotiate
Sriram writes:

>** As far as message integrity goes, yes I am talking about SIP 
>level message integrity. Choosing a baseline algorithm is definitely 
>easier. But then you are defeating the flexibility of SIP (which was 
>one of the original goals of the protocol correct?)

Jonathan's proposal of choosing a baseline algorithm would be the
easiest way. But I'm afraid that as we talk about security, the ability
to change algorithms will be a requirement. One reason for this is
that algorithms could be broken and other ones would have to be
used thereafter. Secondly, algorithms development will introduce
new algorithms with different characteristics, and people want to
take advantage of this. (All security protocols seem to offer this
functionality in one way or another, even, say, HTTP Digest.)

Jari




_______________________________________________
Sip-security mailing list
Sip-security@ietf.org
http://www1.ietf.org/mailman/listinfo/sip-security


From sip-security-admin@ietf.org  Tue Sep 18 12:24:51 2001
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA24531
	for <sip-security-archive@odin.ietf.org>; Tue, 18 Sep 2001 12:24:49 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA23037;
	Tue, 18 Sep 2001 11:50:53 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA22945
	for <sip-security@optimus.ietf.org>; Tue, 18 Sep 2001 11:50:48 -0400 (EDT)
Received: from zrc2s03g.us.nortel.com (zrc2s03g.nortelnetworks.com [47.103.122.66])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA23551;
	Tue, 18 Sep 2001 11:50:45 -0400 (EDT)
Received: from smtprch2.nortel.com (erchg0k.us.nortel.com [47.113.64.104])
	by zrc2s03g.us.nortel.com (8.9.3+Sun/8.9.1) with ESMTP id KAA03642;
	Tue, 18 Sep 2001 10:49:34 -0500 (CDT)
Received: from zrchb200.us.nortel.com by smtprch2.nortel.com;
          Tue, 18 Sep 2001 10:42:55 -0500
Received: by zrchb200.us.nortel.com with Internet Mail Service (5.5.2653.19) 
          id <SK3LNBF1>; Tue, 18 Sep 2001 10:49:17 -0500
Message-ID: <9A9367D1556AD21182C40000F80930AB04411DC3@crchy28b.us.nortel.com>
From: "Sriram Parameswar" <sriramp@nortelnetworks.com>
To: "'Jari Arkko'" <jari.arkko@kolumbus.fi>, sip <sip@ietf.org>
Cc: "jari.arkko" <jari.arkko@ericsson.fi>,
        "Lee Valerius" <valerius@nortelnetworks.com>,
        "tiina.s.koskinen" <tiina.s.koskinen@nokia.com>,
        sip-security <sip-security@ietf.org>
Date: Tue, 18 Sep 2001 10:49:07 -0500
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: multipart/alternative;
              boundary="----_=_NextPart_001_01C14059.73A3DBB0"
X-Orig: <sriramp@americasm01.nt.com>
Subject: [Sip-security] RE: [Sip] SIP Negotiate
Sender: sip-security-admin@ietf.org
Errors-To: sip-security-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Security Issues for the SIP protocol <sip-security.ietf.org>
X-BeenThere: sip-security@ietf.org

This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.

------_=_NextPart_001_01C14059.73A3DBB0
Content-Type: text/plain;
	charset="iso-8859-1"

Jari:

Sorry for the delay - please see inline. 

Regards,

Sriram

__________________________________________
Sriram Parameswar              Phone: 972-685-8540
Interactive Multimedia Server (IMS) Fax: 972-685-3563
Nortel Networks, Richardson USA  Email: sriramp@nortelnetworks.com


-----Original Message-----
From: Jari Arkko [mailto:jari.arkko@kolumbus.fi]
Sent: Monday, September 03, 2001 1:51 PM
To: Parameswar, Sriram [NGB:B658:EXCH]; sip
Cc: jari.arkko; Valerius, Lee [RICH2:2H12:EXCH]; tiina.s.koskinen;
sip-security
Subject: Re: [Sip] SIP Negotiate


RE: [Sipping] SIP Negotiate
I agree that there is a need for this kind of functionality. However,
when you say in the beginning "Examples of parameters that
... include ... message integrity mechanisms, encryption
algorithms, etc.", I'd like to note that this what we need in
terms of security parameters is a _secure_ negotiation of them.
As you say yourself in section 10.1 of
http://www.obsidian97.com/draft-spbs-sip-negotiate-00.txt :

> One of the intended uses for the negotiation mechanism is to set
> up authentication and message integrity algorithms. Thus the 
> NEGOTIATE method itself is not subject to these protections - this
> makes the recipient vulnerable to the Man-in-the-middle attacks. 
> Views are solicited to mitigate this security consideration.

This is a crucial point to handle correctly. Otherwise the peers
may be fooled onto using insecure algorithms or perhaps even
no algorithms at all. It is also interesting to note that the answer
can't really be "integrity protect the NEGOTIATE" because
that would mean an integrity protection algorithm would have
to be applied before one is selected!

[Sriram] One way around this is to ensure that the UA is authenticated
post NEGOTIATE - the authentication algorithm being one of the items
being negotiated. This would diminish the chances of a MITM attack.
Also the message post negotiate may be integrity protected which
is further strengthens security. This of course would require
keys to be sent back in the response to the NEGOTIATE. The
combination of authentication and message integrity would be
good solution to MITM attacks.

We've done some work on this issue earlier in the context of
SIP usage in 3G networks, and came up with some requirements.
These included the following:

* MUST be able to choose different integrity algorithms,
  MAY be able to choose other security parameters,
  encryption algorithms, and security protocols.
[Sriram] Met by NEGOTIATE
* It MUST be possible to protect the parameter negotiation
  against attackers who do not have the secret authentication
  keys (and can't forge hashes in real-time).
* MUST not add new roundtrips (i.e. done in parallel with
  other setup such as registration or authentication). In particular,
  it would not be nice if we had separate negotiation phases for security,
  compression, sip itself, and only then we'd be able to do
  something.
[Sriram] NEGOTIATE does add one more roundtrip however so does sending
in a REGISTER and having a 401 challenge. The advantage of NEGOTIATE
is that several different meta-session parameters can be negotiated
at the same time.

Traditionally, security protocols have solved secure negotiation
of parameters by hashing/signing the offers with authentication
keys or keys derived from authentication. For instance, IKE
negotiates parameters in cleartext but in a later phase a keyed
hash is taken from the original parameter offers and responses,
and sent to the other side as a proof that nothing was changed in
the packet's path.

Something similar needs to happen here as well. We've had
the idea that once integrity protection can be turned on, you'll
repeat the offers under it, and only after this you can start to
trust and use what was negotiated. (Integrity protection can be
turned on basically after something has given you keys
and you have agreed on the selected algorithm.)

Jari Arkko
Ericsson




------_=_NextPart_001_01C14059.73A3DBB0
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; =
charset=3Diso-8859-1">
<META NAME=3D"Generator" CONTENT=3D"MS Exchange Server version =
5.5.2654.89">
<TITLE>RE: [Sip] SIP Negotiate</TITLE>
</HEAD>
<BODY>

<P><FONT SIZE=3D2>Jari:</FONT>
</P>

<P><FONT SIZE=3D2>Sorry for the delay - please see inline. </FONT>
</P>

<P><FONT SIZE=3D2>Regards,</FONT>
</P>

<P><FONT SIZE=3D2>Sriram</FONT>
</P>

<P><FONT SIZE=3D2>__________________________________________</FONT>
<BR><FONT SIZE=3D2>Sriram =
Parameswar&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&n=
bsp;&nbsp;&nbsp; Phone: 972-685-8540</FONT>
<BR><FONT SIZE=3D2>Interactive Multimedia Server (IMS) Fax: =
972-685-3563</FONT>
<BR><FONT SIZE=3D2>Nortel Networks, Richardson USA&nbsp; Email: =
sriramp@nortelnetworks.com</FONT>
</P>
<BR>

<P><FONT SIZE=3D2>-----Original Message-----</FONT>
<BR><FONT SIZE=3D2>From: Jari Arkko [<A =
HREF=3D"mailto:jari.arkko@kolumbus.fi">mailto:jari.arkko@kolumbus.fi</A>=
]</FONT>
<BR><FONT SIZE=3D2>Sent: Monday, September 03, 2001 1:51 PM</FONT>
<BR><FONT SIZE=3D2>To: Parameswar, Sriram [NGB:B658:EXCH]; sip</FONT>
<BR><FONT SIZE=3D2>Cc: jari.arkko; Valerius, Lee [RICH2:2H12:EXCH]; =
tiina.s.koskinen;</FONT>
<BR><FONT SIZE=3D2>sip-security</FONT>
<BR><FONT SIZE=3D2>Subject: Re: [Sip] SIP Negotiate</FONT>
</P>
<BR>

<P><FONT SIZE=3D2>RE: [Sipping] SIP Negotiate</FONT>
<BR><FONT SIZE=3D2>I agree that there is a need for this kind of =
functionality. However,</FONT>
<BR><FONT SIZE=3D2>when you say in the beginning &quot;Examples of =
parameters that</FONT>
<BR><FONT SIZE=3D2>... include ... message integrity mechanisms, =
encryption</FONT>
<BR><FONT SIZE=3D2>algorithms, etc.&quot;, I'd like to note that this =
what we need in</FONT>
<BR><FONT SIZE=3D2>terms of security parameters is a _secure_ =
negotiation of them.</FONT>
<BR><FONT SIZE=3D2>As you say yourself in section 10.1 of</FONT>
<BR><FONT SIZE=3D2><A =
HREF=3D"http://www.obsidian97.com/draft-spbs-sip-negotiate-00.txt" =
TARGET=3D"_blank">http://www.obsidian97.com/draft-spbs-sip-negotiate-00.=
txt</A> :</FONT>
</P>

<P><FONT SIZE=3D2>&gt; One of the intended uses for the negotiation =
mechanism is to set</FONT>
<BR><FONT SIZE=3D2>&gt; up authentication and message integrity =
algorithms. Thus the </FONT>
<BR><FONT SIZE=3D2>&gt; NEGOTIATE method itself is not subject to these =
protections - this</FONT>
<BR><FONT SIZE=3D2>&gt; makes the recipient vulnerable to the =
Man-in-the-middle attacks. </FONT>
<BR><FONT SIZE=3D2>&gt; Views are solicited to mitigate this security =
consideration.</FONT>
</P>

<P><FONT SIZE=3D2>This is a crucial point to handle correctly. =
Otherwise the peers</FONT>
<BR><FONT SIZE=3D2>may be fooled onto using insecure algorithms or =
perhaps even</FONT>
<BR><FONT SIZE=3D2>no algorithms at all. It is also interesting to note =
that the answer</FONT>
<BR><FONT SIZE=3D2>can't really be &quot;integrity protect the =
NEGOTIATE&quot; because</FONT>
<BR><FONT SIZE=3D2>that would mean an integrity protection algorithm =
would have</FONT>
<BR><FONT SIZE=3D2>to be applied before one is selected!</FONT>
</P>

<P><FONT SIZE=3D2>[Sriram] One way around this is to ensure that the UA =
is authenticated</FONT>
<BR><FONT SIZE=3D2>post NEGOTIATE - the authentication algorithm being =
one of the items</FONT>
<BR><FONT SIZE=3D2>being negotiated. This would diminish the chances of =
a MITM attack.</FONT>
<BR><FONT SIZE=3D2>Also the message post negotiate may be integrity =
protected which</FONT>
<BR><FONT SIZE=3D2>is further strengthens security. This of course =
would require</FONT>
<BR><FONT SIZE=3D2>keys to be sent back in the response to the =
NEGOTIATE. The</FONT>
<BR><FONT SIZE=3D2>combination of authentication and message integrity =
would be</FONT>
<BR><FONT SIZE=3D2>good solution to MITM attacks.</FONT>
</P>

<P><FONT SIZE=3D2>We've done some work on this issue earlier in the =
context of</FONT>
<BR><FONT SIZE=3D2>SIP usage in 3G networks, and came up with some =
requirements.</FONT>
<BR><FONT SIZE=3D2>These included the following:</FONT>
</P>

<P><FONT SIZE=3D2>* MUST be able to choose different integrity =
algorithms,</FONT>
<BR><FONT SIZE=3D2>&nbsp; MAY be able to choose other security =
parameters,</FONT>
<BR><FONT SIZE=3D2>&nbsp; encryption algorithms, and security =
protocols.</FONT>
<BR><FONT SIZE=3D2>[Sriram] Met by NEGOTIATE</FONT>
<BR><FONT SIZE=3D2>* It MUST be possible to protect the parameter =
negotiation</FONT>
<BR><FONT SIZE=3D2>&nbsp; against attackers who do not have the secret =
authentication</FONT>
<BR><FONT SIZE=3D2>&nbsp; keys (and can't forge hashes in =
real-time).</FONT>
<BR><FONT SIZE=3D2>* MUST not add new roundtrips (i.e. done in parallel =
with</FONT>
<BR><FONT SIZE=3D2>&nbsp; other setup such as registration or =
authentication). In particular,</FONT>
<BR><FONT SIZE=3D2>&nbsp; it would not be nice if we had separate =
negotiation phases for security,</FONT>
<BR><FONT SIZE=3D2>&nbsp; compression, sip itself, and only then we'd =
be able to do</FONT>
<BR><FONT SIZE=3D2>&nbsp; something.</FONT>
<BR><FONT SIZE=3D2>[Sriram] NEGOTIATE does add one more roundtrip =
however so does sending</FONT>
<BR><FONT SIZE=3D2>in a REGISTER and having a 401 challenge. The =
advantage of NEGOTIATE</FONT>
<BR><FONT SIZE=3D2>is that several different meta-session parameters =
can be negotiated</FONT>
<BR><FONT SIZE=3D2>at the same time.</FONT>
</P>

<P><FONT SIZE=3D2>Traditionally, security protocols have solved secure =
negotiation</FONT>
<BR><FONT SIZE=3D2>of parameters by hashing/signing the offers with =
authentication</FONT>
<BR><FONT SIZE=3D2>keys or keys derived from authentication. For =
instance, IKE</FONT>
<BR><FONT SIZE=3D2>negotiates parameters in cleartext but in a later =
phase a keyed</FONT>
<BR><FONT SIZE=3D2>hash is taken from the original parameter offers and =
responses,</FONT>
<BR><FONT SIZE=3D2>and sent to the other side as a proof that nothing =
was changed in</FONT>
<BR><FONT SIZE=3D2>the packet's path.</FONT>
</P>

<P><FONT SIZE=3D2>Something similar needs to happen here as well. We've =
had</FONT>
<BR><FONT SIZE=3D2>the idea that once integrity protection can be =
turned on, you'll</FONT>
<BR><FONT SIZE=3D2>repeat the offers under it, and only after this you =
can start to</FONT>
<BR><FONT SIZE=3D2>trust and use what was negotiated. (Integrity =
protection can be</FONT>
<BR><FONT SIZE=3D2>turned on basically after something has given you =
keys</FONT>
<BR><FONT SIZE=3D2>and you have agreed on the selected =
algorithm.)</FONT>
</P>

<P><FONT SIZE=3D2>Jari Arkko</FONT>
<BR><FONT SIZE=3D2>Ericsson</FONT>
</P>
<BR>
<BR>

</BODY>
</HTML>
------_=_NextPart_001_01C14059.73A3DBB0--

_______________________________________________
Sip-security mailing list
Sip-security@ietf.org
http://www1.ietf.org/mailman/listinfo/sip-security


