Cristel Pelsser's blog


Distributed Route Servers: A novel internal BGP route distribution architecture

with 2 comments

Lately, I’ve been looking at solutions to avoid correctness issues that may arise in iBGP. These issues are explained in Tim Griffin’s et al., Sigcomm 2002′s paper.

In Pelsser_DRS_IEICE_200903, I proposed a solution relying on an overlay of distributed servers that compute routes on behalf of the routers. Today, each router computes for every prefix its own routes, based on a limited visibility of the routes. In this proposal, we rely on a different distribution of the route computation. Each server has full visibility of the routes for a subset of the prefixes. A server computes for its subset of the prefixes, the route to be used by every router of the domain. The server performs a per-router computation. Full routing visibility at the node performing route computation ensures iBGP correctness.

A presentation describing this proposal can be found here.

Written by admin

May 27th, 2010 at 1:18 am

2 Responses to 'Distributed Route Servers: A novel internal BGP route distribution architecture'

Subscribe to comments with RSS or TrackBack to 'Distributed Route Servers: A novel internal BGP route distribution architecture'.

  1. Cristel,

    I guess this is the concept you explained to me at the IETF 78, right? Do you have any public paper that describes the concept in detail? For example, what paths do you want to provide to every client: Only the best path, several paths, …?

    Anyway, selecting routing information from the receiver’s point of view seems to arouse more and more interest: Now we have my centralized iBGP Route Server architecture, your distributed Route Servers concept, and the Optimal BGP Route Reflection draft of Robert Raszuk et al. I believe that the right direction to solve the iBGP anomaly problem and ensure scalability.


    Uli Bornhauser

    27 Nov 10 at 9:04 pm

  2. Hi Uli,

    Yes, this is the work I mentioned at the IETF. Here is the reference for
    Cristel Pelsser, Akeo Masuda, Kohei Shiomoto. “A novel internal BGP route distribution architecture”. In Proc. of the IEICE General Conference, March 2009.

    The “distributed route server concept” is also used in the following paper:
    Cristel Pelsser, Akeo Masuda and Kohei Shiomoto. “Scalable Support of Interdomain Routes in a Single AS”. In Proceedings of IEEE Globecom 2009, December 2009.
    This paper aims toward a scalable iBGP solution (in addition to correctness). The BGP routes are handled by the Distributed Route Servers architecture that performs per-PoP route selection. Then, the BGP routes are off-loaded to the PoPs. They are stored in a DHT. The routes for a PoP are stored in the PoP, i.e. close to where they will be needed. Routes are then retrieved on demand by the ASBRs. The on-demand part is the tricky bit and needs further work.



    15 Dec 10 at 12:11 pm

Leave a Reply