Another you might want to look at is our system Hubble (http://hubble.cs.washington.edu), which monitors reachability problems. I presented a preliminary study at NANOG40 (http://www.nanog.org/mtg-0706/bassett.html), and now it is running continuously. Some of the other changes since the talk:
- It monitors 78,000+ prefixes, with pings from 100 sites and traceroutes from 35 sites around the world. Once we roll over to the list used by the sister project iPlane, the number of prefixes will be 142,000+.
- We developed a technique using source-spoofed probes to isolate the failures to forward or reverse paths.
- The website lets you query for the status of an address.
For now, failed pings trigger traceroutes, but we plan to start looking at latency changes as well. If you look at the site, I'd appreciate any feedback on what we have and where we should be going with it, as we'd like this to become a useful tool. If you have a prefix we can start monitoring, please feel free to write me or use the feedback form on the website to give us a pingable address to use, and I should be able to incorporate it.
Thanks.
ethan