nanog mailing list archives
RE: Pinging routers for network status
From: Blaine Christian <blaine () inbound blaines net>
Date: Mon, 18 Dec 2000 10:40:04 -0500 (EST)
Hello, One thing to keep in mind is that even if you escape being process switched you still have ECMP and variants to contend with. Some vendors reliably switch packets through certain paths based on source/destination hashes. This means a probe THROUGH a router using a single source address could well report good things while an adjacent source prefix could be screwed by riding a congested path. The only way to be absolutely sure is to either stop using ECMP or run your tests from enough devices to be hashed over all possible paths. All the above in mind, ICMP is quite useful and while it is not 100% accurate it is a pretty good way to identify possible issues. Regards, Blaine
Current thread:
- Pinging routers for network status Sean Donelan (Dec 17)
- Re: Pinging routers for network status Paul Vixie (Dec 18)
- RE: Pinging routers for network status Matt Levine (Dec 18)
- Re: Pinging routers for network status Miguel A.L. Paraz (Dec 18)
- RE: Pinging routers for network status Matt Levine (Dec 18)
- RE: Pinging routers for network status Matt Levine (Dec 18)
- Re: Pinging routers for network status Paul Vixie (Dec 18)
- Re: Pinging routers for network status Bill Woodcock (Dec 18)
- Re: Pinging routers for network status John M . Brown (Dec 18)
- Re: Pinging routers for network status Bill Woodcock (Dec 18)
- Re: Pinging routers for network status Steven J. Sobol (Dec 18)
- RE: Pinging routers for network status Jason Lewis (Dec 19)
- Re: Pinging routers for network status John M . Brown (Dec 18)
- <Possible follow-ups>
- RE: Pinging routers for network status Blaine Christian (Dec 18)