I agree that it shouldn't be precluded. That said, we went to some effort
to put that into OSPF and IS-IS, and did experimentation with it in the
80's. But the commercial importance was nil - Proteon and 3COM implemented
the feature, Proteon found a few takers, but nobody else did that I ever
heard of. Nowadays, the lowest delay route, the highest throughput route,
the least cost, and the most reliable, are all the same path - the optical one.
At 02:54 PM 12/14/2001, RJ Atkinson wrote:
> Something that was useful in the circa late 80s DDN was
>the use of the RFC-791 ToS bits to help in path selection. In
>particular, it was helpful in sorting out delay-sensitive traffic
>(candidate for routing via SATCOM vice cable). This capability
>would still be useful. At the time, it was implemented in
>John Moy's OSPF, which was deployed in those networks (might still be;
>not sure).
>
> It would be nice/helpful if any new architecture did not
>preclude those sorts of routing/traffic optimisations. I imagine
>one would not, but just thought I might mention it in passing
>since not everyone remembers that particular deployment these days.
>
> Just food for thought going forward.
>
>Ran
>rja@inet.org
This archive was generated by hypermail 2b29 : Mon Aug 04 2003 - 04:10:03 EDT