meeting report - comments?

[Scott Brim <[e-mail]>]

Here is a first attempt at a report.  I'd appreciate comments, editing,
etc. on this from those who attended our meetings at IETF.  Eventually
what gets produced will be inserted in the proceedings
						Thanks ... Scott

Topology Engineering Working Group (TEWG)


Current Meeting Report - 2 May 1990

	Agenda

		Report on Europe - Peter Kirstein
		Report on the Pacific - Torben Nielsen
		Report on RFC work - Kent England
		Old and new issues of concern - Scott Brim
		Internet Cartography Project - Ted Bruner 

	Coordinating international connectivity has become a significant
	issue for this group, first because there are few if any
	precedents for what is happening internationally; and second
	because it is all happening so fast.

	Report on Europe - Peter Kirstein

		Peter Kirstein described developments in Europe with
		special emphasis on links to North America.  He pointed
		out the problems managing shared resource "fat pipes",
		multiplexed or not, where different links may have quite
		different use restrictions or resource allocation
		policies and thus, as an example, complex backup
		strategies.  The problems here can't be solved by a just
		a technical or just an administrative group.  "Do we
		know how to manage bits of SPAN separated by bits of
		DARPA?"

	Report on the Pacific - Torben Nielsen

		Torben Nielsen mostly gave a status report on the
		Pacific.  Korea is now on.  At some point New Zealand
		will be daisy-chained to Australia and the direct link
		to Hawaii will be removed.  Japan has multiple medium
		speed links; working on merging them.  Taiwan soon.
		Australia is no longer urging colored book protocols.
		Link to Europe within a year.  Singapore, Hong Kong,
		Indonesia are talking; Thailand and Malaysia are
		interested.

	Report on RFC work - Kent England

		Progress is being made on the RFC for generic mid-level
		routing policy and "rules of thumb".  New guidelines
		presented at the meeting were: explicitly engineer every
		fallback -- none should be accidental; avoid routing
		"ties" -- there should be distinct preferences, to avoid
		bistable situations; and the hardest problem to diagnose
		is oscillation.  The group working on the RFC continued
		that night.

	Old and new issues of concern - Scott Brim

		We had to skip this section because we were out of time.
		We are continuing on the TEWG mail list.

		Items that have been resolved since the last meeting:

			VMNET interaction with the Internet: VMNET is
			still being planned; they will now be more
			conscious of the physical topology of the
			Internet when designing their traffic flows.

			CSNet transcontinental link: Dan Long has
			written a routing plan which demonstrates how
			they are being careful not to cause routing
			problems with this link.

			CA*Net and its multiple connections to NSFNET
			(and NASA): Dennis Ferguson has written a plan
			describing how they will use their multiple
			connections.

		Items that have been brought up outside the meeting so
		far are:

			The Army Supercomputer Network, and how it will
			interact with the rest of the Internet.

			Paths which have both ends in the United States
			but *unintentionally* travel through other
			countries.  This same problem exists for other
			countries as long as they have possible fallback
			paths through other countries.

			NASA's ACTS satellite system and how it will
			interact with the Internet.

	Internet Cartography Project

		Worked on jointly with the NJM working group, and
		presented in their report.

	Attendees