On Thu, Mar 26, 2009 at 10:46 AM, Robert Raszuk <[email protected]> wrote:
> Hi Lixia, > > Hi Robert, >> >> BMP is about what data to suck out from a BGP peer >> this work is about what's the best format to put that data in >> > > BMP defines what to send and how to send it ... it defines a new header to > be used as a wrapper for the "raw" data. > As such it arrives on the management station. > > What would such station do with it is outside of scope of bmp draft, but as > currently defined/implemented there is no additional encoding of it between > the router and mgmt station what so ever :). > True, BMP encapsulates bgp 'raw' data with a simple header. In such header, there is timestamp, peering information which is not in bgp segments but is useful and necessary for the management station. > > Hence my question to clarify overlap between this work and bmp. > As you said, BMP defines what to send (Adj-RIB-in, Loc-RIB, Updates) and how to send it(wrapped in a header over TCP). In this draft, we focus on HOW to faithfually represent BGP segments in XML, while also providing flexibility for annotation or taging. We think It would be better to have a standard xml schema as early as possible before multiple parties develop their own imcompatible xml formats. This schema is not finallized and open to change based on working group comments. As for BMP, what maybe overlapped with bmp here is not this draft, but the BGPMon project which we use to generates xml bgp messages. The major difference is that a BMP process "resides" in a bgp router, so it is capable to dump the whole adj-rib-in, while our BGPMon "peers" with routers, so it only gets the best-path. Payne > > Cheers, > R. > > > > _______________________________________________ > GROW mailing list > [email protected] > https://www.ietf.org/mailman/listinfo/grow >
_______________________________________________ GROW mailing list [email protected] https://www.ietf.org/mailman/listinfo/grow
