The observations suggest something different. Remember that I mentioned that this blockage only occurs for a SINGLE IP in my network: the base IP belonging to the server doing the stats collection every 5 minutes. Other IP's (even on that server) are ok. Rate limiting would be visible as packet loss after the committed rate is exceeded: you can't rate-limit on production border routers down to single-host granularity, only a big interface-wide rate-limit will work. I am still trying to determine if the interface IP in question is the customer side of things, but things are a bit murky there, because there appear to be ACLs on both sides of the link. At Wednesday 04:15 PM 6/14/00 , Mufti Ahmed <Mufti.Ahmed@reuters.com> wrote:

Kai what if UUnet is running some type of QOS mechanism in their core network that prevents massive amounts of icmp messages traveling through? Maybe they are rate-limiting everyone with CAR. Their infrastructure is interesting they use Cisco, Fore, and Juniper at least in production. And they claim they run an MPLS network; maybe this is where icmp traffic is killed..