Cook’s Collaborative Edge

From my mail list as excerpted by Tim Shepard from another post:

“A new IRTF research group, End-Middle-End (EME) Research Group, has been created with an appended charter. Too clarify, don’t get too excited about possible change. This is the IETF “research group” and it can take a long time for anything to happen.”

Tim Shepard (BS and PhD from MIT - [email protected]), who is interested in large complex communications systems and has been consulting and contracting for the past 8 years,, responded with something that to me cuts through all the smoke, fire, and brimstone about whether IPv6 will ever be successfully deployed.

Tim said: It’s in the IRTF, an organization that was created at the same time as the IETF and exists alongside the IETF. IRTF != IETF. Think of the IRTF as a meta-club of loose research clubs.

The IETF has been trying to figure out how to cope with the erosion of the end-to-end model for 10 years or more, and this particular development is nothing new. I remember the contentious meetings regarding what should be done about NATs at the IETF meeting in Munich in 1997. Since then, much has happened, and the debate about whether protocols should be redesigned to work well with NATs or NATs should be redesigned to work well with protocols continues.

I hesitate to provide any pointers to particular goings on for it would be like pointing out some particular square-foot of the elephant’s hide. One relevant WG currently underway is http://www.ietf.org/html.charters/behave-charter.html

If you want to see what the EME RG might be about, see
Handley’s “Connection Signaling” slides from the Offpath BOF in Montreal.

I’ve believed for awhile now that the best way we might solve these problems is to seperate the ID and locators which today are wrapped up in what we call an “IP address”. HIP and Shim6 are the two relevant goings-on there.

The IETF moves slowly. I doubt anything else could move any quicker and actually accomplish anything for the Internet at large.

IPv6 is here already.

As someone at the IETF meeting last week mentioned to me: Only in the English speaking world does anyone think that IPv6 has not yet happened, e.g. anywhere there are Internet cafe’s catering to a significant Korean expat clientele will have gotten their IPv6 working so that the Korean expats can get to the IPv6-only web servers that have content they want to read.

I got my machines at home live on the global IPv6 network four years ago using RFC3068 and RFC3056 which did not require any cooperation from my ISP. More recently I installed Miredo (an implementation of Teredo for Linux) and now I have working IPv6 connectivity even when I’m behind IPv4 NATs wherever I may be.

Vista will have IPv6 turned on and connected one way or another. It includes an implementation of Teredo and will use it if necessary.

IPv6 can happen in a big way in the next few years and the old ISPs may be the last to know. The users will move first. Some years or decades later, people will stop caring about whether they have IPv4 connectivity, and IPv4 will whither away.

Cook’s Edge: I had a voice chat with Tim to get his permission to post. His strong recommendation to all readers was to grab and read the next two urls. Tim added that the last place he ever expects to see IPv6 deployed is in large commercial operational IPv4 backbones - saying essentially it will become pervasive but tunneling everywhere and this tunneling is already happening.

Tim: If you want to understand the technical challenges facing the plumbing of the net and the historical context, I suggest:

Wither Routing, by Geoff Huston and

Why the Internet only just works by Mark Handley.

Trackback URI | Comments RSS