[e2e] [Fwd: State of MPLS deployments today]
David P. Reed
dpreed at reed.com
Fri Oct 4 12:20:17 PDT 2002
At 08:18 AM 10/4/2002 -0700, Christian Huitema wrote:
>Computing path at the edge of the network is intellectually tempting:
>pushed to the extreme, it gives you the perfect "stupid network"
>architecture. However, doing so bumps into three important obstacles: do
>the edge points have enough information to compute the path correctly?
>Can the edge points synchronize their decision effectively? How do the
>owners of intermediate resource control the usage of their property?
This is exactly the right way to think about this, IMO.
>It is not surprising that MPLS ends up targeting a narrow application,
>basically a form of traffic engineering. Answering the three previous
>questions gets much simpler when the "edges" are the edges of a provider
>network, rather than the edges of the Internet, and when the timescale
>over which the decisions are made is expressed in days or months, rather
>than seconds or milliseconds: a provider can obtain information on
>network usage by statistical analysis of traffic over a long period; the
>edge points can be coordinated by a form of central planning; and the
>decisions are made by the very owner of the resource.
One is, of course, left with the question of how much it helps, even in
this application. That people buy it and deploy it is necessary, but not
sufficient. People bought the V-Chip. Don't know if more than 1% of the
people who bought it, use it for anything, though it is "used" by
everybody. People bought ISDN, but they used it for high-speed bit
transfer, not AIN telephone services (did anybody use the D-channel?)
More information about the end2end-interest