CURRENT MEETING REPORT


Minutes for Internet Stream Protocol V2 Working Group (st2)

Reported by: Tim O'Malley, BBN Corporation and Lou Berger, BBN 
Corporation


The Working Group met for one session in Dallas.  Lou Berger chaired 
the session.  Tim O'Malley acted as scribe.  The main focus of the session 
was the State Machine Internet-Draft, draft-ietf-st2-state-00.ps.  The 
agenda for the session was:

o  Status of group
o  Implementation status
o  State machines discussion
o  Open Issues
o  Wrap-up

There was also an administrative announcement.  The Working Group 
mail and FTP server is being shutdown.  The new server is located at 
BBN.  The new server information is:

o  Mail List:		st@bbn.com
o  Request Address:	st-request@bbn.com
o  FTP Server:		ftp://ftp.bbn.com/pub/st

The status of the group was reviewed.  The Working Group has 
accomplished the main chartered objective of publishing a revised 
protocol specification.  The specification, RFC 1819, was issued over the 
summer.  The only remaining work is the completion of the State 
Machine document.  The state machine document has taken much longer 
to prepare than anticipated, but good progress has been made.  Since 
the Working Group has completed its main chartered objective, and the 
technical work on State Machines is mostly complete, it was announced 
that the Working Group will close before the next IETF in March.  
There were no objections from the Working Group.

The status report was followed by implementation reports from GMU 
and NTT.  Professor Mark Pullen reported on a ST2+ Prototype being 
worked on at George Mason University.  This work is an unsponsored, 
independent study project being conducted by several Masters Students.  
The goals of the project were to verify and implement ST2+ SCMP, and 
to define and implement a service (not ST specific) API.  The project did 
not implement neighbor failure detection (HELLO), and used IP 
encapsulation.  The work is partially complete, and is expected to be 
fully functional by February 1996.  Professor Pullen reported that the 
most surprising aspect of the implementation was the amount of code 
needed to implement the API and the protocol. (The breakdown 
between the two was not known, but can be obtained from GMU.)  The 
code is written is user space, and runs with Solaris on a Sparc .

Muneyoshi Suzuki from NTT Telecommunications Networks Labs 
reported on an effort to build an ST2+ over ATM system.  The work is in 
a planning phase and protocol design phase.  They plan to have a 
finished implementation in 2 years.  The work is focused on running over 
VBR and CBR, and mapping SCMP into Q.2931.  The biggest difficulty 
encountered has been mapping the ST2+ FlowSpec into ATM.  Mark 
Pullen commented that he too has seen this problem, and will raise the 
issue in the INT-SERV Working Group when appropriate.

Sharon Sergeant, one of the State Machine document authors, then 
reviewed the Internet Draft.  She gave an overview of the document, its 
structure, how the agent functions s map into state machines.  She also 
reviewed the Retry and Control FSMs in detail.  She reported that the 
main work remaining one the document is:

o  Language clarification 
o  Minor technical corrections
o  Reorganization of document for clarification
o  Removal of implementation specific discussions

She also solicited help in completing the draft.  Mark Pullen offered to 
look for a student to help.  A few people offered to review the 
document.

The Working Group closed with a review of the decision to close the 
group by the next IETF and a review of action items for the draft.  The 
next version of the document will include a major reorganization and 
minor technical corrections.  It will be issued in mid-January.  Then, 
based on feedback, another draft will be issued in mid-February.  The 
desire is for this draft to be the final draft, and for it to be published as 
an Informational RFC by mid-March.