Chicago IETF (August, 1998) IPsec Working Group Meeting Minutes

The WG met on Tuesday at the IETF meeting in Chicago, from 14:15 to 15:15.  
Approximately 120 people attended.  This was MBONE broadcast.

The Agenda was: 

* Workgroup status
* Workshop announcement
* Charter revision
* Discovered problems with Ipsec/IKE based on current implementation  experience --- 
   Lifetime discussion
* ICMP messages, standardized error codes, and MIBs
* Policy/tunnel endpoint discovery
* Policy-based Security Management

Workgroup status

Ted gave a report on the status of the IPSEC working group.  The full suite of Internet 
Drafts have been approved by the IESG.  They are currently being processed by the 
RFC editor and should be published shortly.  It is now time to revisit the IPSEC charter 
since we have met almost all of the goals and milstones in the original charter.

Workshop announcement

Microsoft will be sponsoring an IPSEC interoperability testing workshop in Redmond 
on Aug 31 -- Sep 3rd.  Approximately 20-25 companies have signed up for the workshop.  
William Dixon (wdixon@microsoft.com) is the contact person for this workshop.

IBM is also sponsoring another round of interoperability testing in Binghamtom, NY 
on October 27--30.  This test will also include L2TP.  The $300 fee has been waived 
by IBM.

Charter revision

Bob Moskowitz led a discussion on new items for work goals for revising the charter.  
These items included:

* address errors and improvements; errata to be maintained on MIT web site
* add functionality
* remote client support
* policy and tunnel endpoint discovery (Bill Simpson votes to remove, since this is a 
   complex non-security issue already being dealt with elsewhere) 
* complex tunnel management
* ICMP messages, error codes, MIBs
* Additional algorithms 
* IPsec over non-IP protocols (IPX? "Running shoes?")
* Key recovery: excluded to the sound of cheers and hisses
* Integration of IPsec and certificate frameworks, DNSsec
* Cleaning up the host-host (non-VPN) case; not sure what's missing
* results implemented from Secure Multicast IRTF activity

The working group chairs will compose a new proposed charter based on these 
suggestions, and present it to the working group.

Lifetime discussion

Although there is a default value established for Phase 2 lifetimes, there is no similar 
default for Phase 1 lifetimes.  There is unfortunately is conflicting interpretations 
regarding how to proceed in the absence of an explicitly specified PHase 1 lifetime.  
There is an optional notification facility, but it's unclear what happens if the notified 
value ins't acceptable.  There is an interoperability impact caused by this 
underspecification.  (This will need to be corrected in a protocol errata.)

ICMP messages, standardized error codes, and MIB's

Michael Richardson gave a presentation on a two problems which he has been 
concentrating on.  One is the issue of Path MTU discovery across a IPSEC tunnel; 
which could be ignored in IPV4, but not in IPV6.  (Since IPV6 drops packets greater 
than the MTU, instead of fragmenting them; on the other hand, the IPV6 minimum 
MTU is also much bigger, so perhaps the problem can be ignored).  

The other area of concern is ICMP messages and IPSEC, to support diagnostic tools 
such as traceroute and PING.  

Policy/tunnel endpoint discovery

Roy Perieira has some drafts forthcoming which will cover IPSEC policy and tunnel 
endpoint discovery issues.  These include how a new machine on the network 
bootstraps itself by obtaining  its first policy, and secure route discovery in the face of 
complex topologies and multiple secure paths for load-balancing and/or redundancy.

Policy-based Security Management 

Luis Sanchez gave a presentation on some Security Policy Management going on at BBN.  
(Slides included)