|
Thread Index Re: [isp-bgp] TE practices when peered to two big upstreams?
On (07/01/05 22:36), Erik Haagsman wrote: > > On Fri, 2005-01-07 at 21:00, joshua sahala wrote: > > > bgp picks the oldest path, all other things being equal, because that > > implies a more stable path. unless bgp is picking the very wrong path, i > > personally don't like to muck with things like this > > Biggest problem with BGP is the fact that AS path is hardly a very good > qualitative metric. Getting it right quality wise (lowest latency, > lowest packet loss) means a LOT of hand-tuning and comparing traces etc. > Doesnt get much better although there's a few BGP optimizing appliances > out there like RouteScience that are supposed to do much better, but at > a high cost and prolly not suited for all situations. it is good enough to be part of the overall path selection, and granted, we both know that not all as'es are as good to transit as others, but colohost was asking about equal length as-paths, so then you have to tweak a different knob (or several of them) the question that i have is why do people want to see a totally even (or as close to it as possible) distribution of traffic on their links, especially when it will take messing with things that probably should not be touched unless absolutely necessary? in my network, i like stability and predictability...'routing magic' boxen tend to make things much less stable and predictable (but then the traffic graphs have a nice flat-line) /joshua To unsubscribe via postal mail, please contact us at: Jupitermedia Corp. Attn: Discussion List Management 475 Park Avenue South New York, NY 10016 Please include the email address which you have been contacted with. Copyright 2005 Jupitermedia Corporation All Rights Reserved.
Thread Index |
|
