Yes, it's a ghastly protocol, and the bgp4mibv2 will cause all sorts of horrible contortions. But the alternative (screen scraping) is worse. Also it's got the advantage that my current monitoring systems support it, so monitoring my bgp sessions is just a matter of changing a few lines in my provisioning system, rather than writing support for a completely new protocol like bmp.
Nick Sent from my iWotsit. On 16 Dec 2010, at 16:35, Jeffrey Haas <[email protected]> wrote: > Nick, > > Addressing one specific point: > > On Wed, Dec 15, 2010 at 11:16:06PM +0000, Nick Hilliard wrote: >> 2. I would much prefer to see support for bgp4mibv2 on regular routers, >> rather than a completely new tool (although I do acknowledge the >> necessity of being able to peek into ribs, and that snmp is woefully >> inadequate for this job). > > Speaking as the editor of the MIB and being known for being opinionated > about SNMP: SNMP sucks. SNMP is *not* a good tool for monitoring routing > state. It's fine for monitoring peer counters and monitoring specific > prefixes using standard polling techniques. > > The MIB also has the ability to let you examine multi-rib scenarios as well, > but the indexing mechanism will have some amount of vendor-specificness to > it. > > -- Jeff _______________________________________________ GROW mailing list [email protected] https://www.ietf.org/mailman/listinfo/grow
