In einer eMail vom 21.12.2008 15:38:19 Westeuropäische Normalzeit schreibt [email protected]:
I think that the real problem is that with current techniques, there is a serious mismatch between the number (millions or in the future, perhaps hundreds of millions or billions) of end-user networks which want multihoming, TE and/or portability and the only way this can currently be provided - by globally routed prefixes for every such end-user network, which every DFZ router needs to handle in its RIB and generally in its FIB. My saying since 45. I have referred to two other networks which both are 1000 times larger: the road network, and the postal service network. None of them has this crazy problem because none of them does user reachability information dissemination. Let me point to a 3rd network: the railroad network. There a single link (=railway between two distant cities) may take 10 years building it. So there is, broken down to a single second, in principle zero change. But if you assume 6 billion people on earth and each of them may move every ten years to another location, and if you disseminated their individual change of address information to all railway stations you will create a considerable update churn. Isn't it very obvious that this concept is absurd ? And that the scalability problem is NOT due to expressing three different concepts, as is stated? It is not open to us to say the problem lies in millions of end-user networks wanting multihoming, TE and/or portability. So the problem is that the current architecture can't scale to meet this demand. That is the guts of the problem. It is only one interpretation of the problem to see it in terms of IP addresses being used for too many purposes all at once, which is your page's current definition of the root cause of the scaling problem. There are potential host-based solutions - with new protocols and changes to stacks and applications - which pushes the responsibility for multihoming, TE and portability to each host, by way of separating out the various functions into physically separate entities, such as for GUID, SID and LOC. But that doesn't mean the only or best definition of the problem is in these terms. Such a solution necessarily involves every host in fussing around with changes in the core of the Net. I see many problems with this: Fundamental objections to a host-based scalable routing solution _http://www.irtf.org/pipermail/rrg/2008-November/000271.html_ (http://www.irtf.org/pipermail/rrg/2008-November/000271.html) Routing is first of all a network issue, but this should not prevent that hosts can take influence on routing wrt their sent/ sent-to packets. Why EITHER - OR ? Why not enable BOTH ? TARA requires such little data, that multi-homed users could -technically- get them from each adapted ISP. Politically, is a different story. A further note about "steal of service/ discriminating routing / internet economic model": I have my doubts that the current internet economic model enables that a transit network in the middle (and not at the ends) of some path can discriminate the packet flow just as well as when being at its end. It seems to me that a special aspect is picked for argumenting against strategies B and C. Another note concerning strategy B: It looks like PNNI, which I do know very well, which I once supported very much, which I tried to emulate at the beginning of NIRA, too, but which does topology aggregation in a wrong way. Heiner
_______________________________________________ rrg mailing list [email protected] https://www.irtf.org/mailman/listinfo/rrg
