In article <hot.mailing-lists.nanog-199611181859.NAA09628@merit.edu>, Craig Labovitz <labovit@merit.edu> wrote:

the larger providers. With the exception of Sprint, most providers seem to have ~10% error in their BGP announcements (of course, this is from a very small sampling).

Sprint install folks have this tendancy to put in static routes regardless of whether you're peering or not. At least that's been my experience getting multi-line setups installed. I'm only half-kidding when I suggest telling a sprint installer "yeah we have our own IPs, 172.16/16". If you do end up peering then they seem to put a static in for the class C that contains your loopback interface. You can get them to change it, but the default seems to be static-everything. Do any providers reserve ips for use on loopbacks? i.e. nets they divide into /32s to point at loopbacks on the customer router. This would seem to be a prudent measure for many multihomed customers -- I'm loathe to use any PI addresses for loopback-peering because of the potential for mistakes with static routes to the loopback address. Dean