Hi Young, Thanks a lot for the feedback. I totally understand your point. We have some internal stuff that needs to be changed in addition to Quagga fix. I'm adding it into our ToDo list. Whenever we will have time, we will look into it. In general, if you will have any other comments about the RIS services, we will be happy to hear it. Regards. Arife On Tue, 8 Nov 2005 17:58:21 -0800 (PST), Young Hyun wrote: * Actually, I'll most likely use RIS data as well as RouteViews data. * However, for my immediate needs in exploring IPv6 BGP tables, RouteViews * is a little more convenient since I can simply go to one collector, * route-views6, and get 2 1/2 years of IPv6 (and only IPv6) snapshots from * up to 13 peers. In contrast, IPv6 tables are more scattered in the RIS * collection. Specifically, the lack of a definitive IPv6 collector having * a high concentration of IPv6 tables makes working with RIS data less * convenient. For example, bview.20051104.0000 from rrc12 has 1,693,790 * records, but only 4,070 records (0.2%) are for IPv6. Also, of the 9 IPv6 * peers of rrc12, only 6 provide any useful amounts of data (the other 3 * peers export 2 or fewer prefixes). So, if I was to use RIS, I'd want to * aggregate IPv6 data across multiple collectors in order to gather as many * peers as I can. But because the bulk of the RIS data files is IPv4 data, * I'll have to spend more time downloading files and weeding out IPv4 data * than I would if I simply use RouteViews data. Admittedly, other * RouteViews peers, such as route-views.wide, also have small amounts of * IPv6 data mixed in with a large amount of IPv4 data, and to be * comprehensive, I ought to download and process that data as well. * * The RIS data sets are valuable, please don't misunderstand me. But it * would make data processing easier if the IPv6 data were, for example, * provided in separately downloadable files--maybe they could be called * bview6.<date>.<time>.gz and be found in the same directory as the current * full tables (having both IPv4 and IPv6). I'm not sure that the IPv6 * updates need to provided separately. This, of course, depends on whether * demand for IPv6 data (separately from IPv4 data) is high enough to justify * the effort (though maybe the effort required is small, requiring a simple * tweak of your Zebra export scripts to create the additional files). * * --Young * * On Tue, 8 Nov 2005, RIPE NCC RIS wrote: * * > Hi Young, * > * > I see you use Route Views data. Just wondering whether you have any * > reasons not to use RIS data. We like to improve our data collection * > process. Your feedback will be quite useful for us. * > * > Regards. * > Arife -- Arife Vural SED, RIPE NCC