|
|
|
|
|
|
|
|
|
|
|
|
| home | help | back | first | fref | pref | prev | next | nref | lref | last | post |
From: Deepak Jain <deepak@ai.net> To: "nanog-post@rsuc.gweep.net" <nanog-post@rsuc.gweep.net>, "nanog@nanog.org" <nanog@nanog.org> Date: Tue, 4 Nov 2008 17:16:35 -0500 In-Reply-To: <20081104211508.GA42816@gweep.net> Errors-To: nanog-bounces@nanog.org > I seem to remember that an AS Path wouldnt be valid with the same AS > appearing more than once in non-sequential order, i.e. at the beginning > and the end in this case, but am unable to verify since everything I've > been reading up on discusses BGP in single-site terms. Your largest concerns will be - how do intend your split sites see each other? GRE/IPSec tunnels are a popular if potentially fragile answer. - deaggregating is antisocial, and depending on your allocation size & in what part of the bit-spectrum it lies, you may need to be concerned about the deaggregates getting filtered. Assuming a decent answer to the first point and you can 'safely' announce the aggregate in both spots to attrach traffic beyind the propagation distance of your deaggregates. You may wish to consider having the same set of providers in each location and examining approaches that make your interior appear more coherent than it is [eg deaggregates with sane MEDS and no-export to allow you to TE across your paid links, and only let the aggregate out past their borders]. --- That is really a great, concise answer. If you don't know enough about BGP = to understand everything Joe is saying here, you may need help doing it right. Basically what your risk factors are is even if someone said "yes, this is = fine." It will probably break somewhere and under some circumstances. And if someo= ne says "no, this won't," you are still faced with joe's first "largest concerns". Deepak
|
|
|
|
|
|
|
|
|
|
|
|
| home | help | back | first | fref | pref | prev | next | nref | lref | last | post |