> Anyone know *which* RFC says a packets should be routed using > the most specific route in a routing table, not (for instance) > the first route in the table that matches, or, for instance, > using a less specific route that has a better metric? It might be worth your while to look through the sections of RFC 791 and subsequent revisions which define the second byte of the IP header, which sets the delivery-priority and the class-of-service for the packet. The class-of-service (type-of-service?) field lets you specify whether you want the shortest, largest, cheapest, or most reliable route, those being mutually exclusive. In my experience, both fields are set to 0 on just about all packets, which doesn't specify any requested class-of-service. That presumably leaves the router free to choose any route it likes, at its discretion. > This is so basic I hardly know where to find it If the router, in its discretion, chooses to route the packet in the wrong direction, that should be easily demonstrated and reasonably incontrovertable, no? What's your vendor's argument in favor of shipping the packet to a nearby site which isn't associated with your destination? I'm kind of curious to hear it... Who's the vendor? Please humiliate them publicly. :-) -Bill ______________________________________________________________________________ bill woodcock woody@zocalo.net woody@nowhere.loopback.edu user@host.domain.com
> This is so basic I hardly know where to find it
Firstly thanks to all who replied.
If the router, in its discretion, chooses to route the packet in the wrong direction, that should be easily demonstrated and reasonably incontrovertable, no? What's your vendor's argument in favor of shipping the packet to a nearby site which isn't associated with your destination? I'm kind of curious to hear it...
It's a host with 2 i/fs not a router to start off with, not that that makes much difference. Their argument is "we wrote the kernel that way and don't want to change it". My customers argument is "I don't want to redesign my network so one of you fix your routers / kernel".
Who's the vendor? Please humiliate them publicly. :-)
They get 1 more chance 1st :-) Alex Bligh Xara Networks
also, 1058, rip RFC 'Entities that use RIP are assumed to use the most specific information available when routing a datagram.' susan Bill Woodcock wrote:
> Anyone know *which* RFC says a packets should be routed using > the most specific route in a routing table, not (for instance) > the first route in the table that matches, or, for instance, > using a less specific route that has a better metric?
It might be worth your while to look through the sections of RFC 791 and subsequent revisions which define the second byte of the IP header, which sets the delivery-priority and the class-of-service for the packet. The class-of-service (type-of-service?) field lets you specify whether you want the shortest, largest, cheapest, or most reliable route, those being mutually exclusive. In my experience, both fields are set to 0 on just about all packets, which doesn't specify any requested class-of-service. That presumably leaves the router free to choose any route it likes, at its discretion.
> This is so basic I hardly know where to find it
If the router, in its discretion, chooses to route the packet in the wrong direction, that should be easily demonstrated and reasonably incontrovertable, no? What's your vendor's argument in favor of shipping the packet to a nearby site which isn't associated with your destination? I'm kind of curious to hear it...
Who's the vendor? Please humiliate them publicly. :-)
-Bill
______________________________________________________________________________ bill woodcock woody@zocalo.net woody@nowhere.loopback.edu user@host.domain.com
-- ============================================================== Susan Roycroft (formerly Susan Holdridge) cisco Systems, Inc. Customer Engineering ccie# 2322 ==============================================================
participants (3)
-
Alex.Bligh -
Bill Woodcock -
Susan Roycroft