For the parser part, you might consider looking at the UCLA bgpparser to see how they dealt with certain issues.
http://irl.cs.ucla.edu/software/bgpparser.html For path questions, I'm guessing you might want to be looking at route leaks. If that is the case, https://ripe69.ripe.net/presentations/157-RIPE-69-Routing-WG.pdf https://ripe69.ripe.net/archives/video/10076/ The related definitions in GROW: http://tools.ietf.org/pdf/draft-ietf-grow-route-leak-problem-definition-00.pdf /jgk On 3/4/15 5:17 PM, Yasuhiro Ohara wrote: > > Hi, > > I'm developing a new tool that can read routeviews > RIB files (MRT v2 RIB format). I call it "bgpdump2". It is similar, but > with a few additional features, to the libbgpdump. > It is going to be open source (the process is on-going). > > This e-mail is to consult if there is some interest in such a tool. > I'm considering to provide a presentation in Dallas. > > The motivations of creating a brand new tool are below. > 1. libbgpdump seemed to have some rooms to speed up, since > it seems to issue many small-sized read operations, rather than > making a small number of large-sized read operations. > 2. I wanted to add a few additional features (explained below) > to such tools. > 3. it didn't seem easy to modify the libbgpdump for me. > > The additional features include longest prefix matching using a > PATRICIA table inside the command, diff'ing the prefixes > between peers (Note, the BGP prefixes are stored per peer > in the RIB file), and to display of some statistics of RIBs per peer. > > I am from a R&D division of a tier-1 ISP (i.e., NTT Communications), > and was trying to compare BGP full route routing table instances, > possibly between ISPs. I expect that such a tool can contribute to > get a detailed view on how/why BGP full route routing tables are > different. For example, we're going to need the longest matching > lookup function to see if the differentiating prefix is covered by a > shorter prefix. > > However I'm not greatly succeeding so far, since the reason for > a prefix that is differentiating depends on case by case > (and there are various reasons including "the business relationship"), > and I could not get a summarized view after the comparison of > (say) two BGP full route routing tables. > > I would like to get some feedback from the community on these things. > I can provide in Dallas some more details of my tool, > some performance benchmark of it, > and some statistics of RouteView's RIB files. > > Would you kindly share your opinions on this ? > > Thanks. > > Best regards, > Yasu > > > _______________________________________________ > GROW mailing list > [email protected] > https://www.ietf.org/mailman/listinfo/grow > _______________________________________________ GROW mailing list [email protected] https://www.ietf.org/mailman/listinfo/grow
