Thanks Jon,
There are a couple of comments I would like on you comments though:
>- DNS
i think dns is subsumed in management - i would say just routing & Management (but possibly also IPv7 evolution).
The whole issue is of course to maintain reasonalbe full control of routing and mangement so that any migration to IPv7, CIDER TUBA etc MAY be possible.
IPv7 is CLNP on the 'internetworking' layer, no magic, and it still have some of the old source/dest limitation's, even thus it might do some good. (Personally I think we need another kind of toy here) Cider, is something we need today, and put's limitations on the topology if we want it to make the big routing table savings we baadly need, NOW! There is somthing more, routing of troble-tickets, with todays international cooperation among all the actors (maybe not EMPB) we can receive a user complain't and get it resolved , even if it crosses *MANY* operators/networks.
As for DNS, some think they can build Internets without DNS support... This is the best single reason why all open Internet service providers has to interconnect. Thus you end up in interconnection agreements, topology and routing issues.
--mats
Not mentioning those who have 'fake' roots that sometimes poisions the whole system for others. This is actually a crusial point, the DNS system is based on countrycodes and it should be made avaliable for all networking organisations in that country, and it has to be fully interconnected. -Peter
The whole issue is of course to maintain reasonalbe full control of routing and mangement so that any migration to IPv7, CIDER TUBA etc MAY be possible. IPv7 is CLNP on the 'internetworking' layer, no magic, and it still have some of the old source/dest limitation's, even thus it might do some good. (Personally I think we need another kind of toy here)
IPv7 may be CLNP or PIP or IPAE or other...it is NOT decided til at earliest next ietf...
There is somthing more, routing of troble-tickets, with todays international cooperation among all the actors (maybe not EMPB) we can receive a user complain't and get it resolved , even if it crosses *MANY* operators/networks.
agree definitely!!! jon
IPv7 is CLNP on the 'internetworking' layer, no magic, and it still have some of the old source/dest limitation's, even thus it might do some good. (Personally I think we need another kind of toy here)
IPv7 is not CLNP. IPv7 is the generic name for the "next version of IP". Current version number in use is IPv4; number 5 and 6 were assigned long ago to experimental variants of the IP protocol; number 7 is thus the next version number, full point. What will be the actual specification for this next version number is still a matter of heated debate. Two prominent proposals are TUBA and IPAE; another candidate is PIP. TUBA is a derivative of CLNP; IPAE and PIP are not. The IAB, in a draft statement issued in June after meeting in Kope, favored TUBA. This raised a storm of hate mail, and the draft statement was retired; it was not published as RFC. I do not think that we will end up with anything like a CLNP based solution for the next generation. CLNP addressing plan is bogus, and the CLNP encodings are, to say the list, suboptimal. You are right to mention that we need "another kind of toy"; promising ideas are appearing in the ongoing debate. And you are also absolutely right when you say that CIDR is the only game in town now. Christian Huitema
participants (3)
-
Christian Huitema -
Jon Crowcroft -
Peter Lothberg