So, I know people use big web sites or DNS servers for ping target but I don't think it's best because of two reasons, first, it seems unethical to use someone's resource that has nothing to do with the service they are providing. And second, those sites could one day just decide not to respond to ping.
I think we are going to set aside some IPs on our network and configure hosts to ping for those routers. Thanks all for the suggestions! On Wed, Oct 20, 2010 at 10:41 AM, David Freedman <[email protected]> wrote: > Jay Nakamura wrote: >> When you use IP SLA to track if an upstream is working on a ISP >> connection (From customer point of view, and you are not the ISP that >> knows what will be safe to ping), what do you usually configure to >> ping? I have found that one hop up from the CPE is not necessary >> reliable on DSL/Cable. I was wondering if anyone can share their >> experience on what works well and what to look out for. > > Dont know about your ISP , but we provide customers with an anycasted > "sponge" which accepts ICMP and various other standard RTR operations, > for this purpose. > > We deployed this mainly as an alternative to running routing protocols > over customer circuits over which there was no link loss detected when a > break in the middle (mainly badly setup 3rd party pseudowire circuits), > am now looking at BFD backed up static routes to replace this. > > Routing protocols are always preferable, IMHO. > > Dave. > > -- > > > David Freedman > Group Network Engineering > Claranet Group > > _______________________________________________ > cisco-nsp mailing list [email protected] > https://puck.nether.net/mailman/listinfo/cisco-nsp > archive at http://puck.nether.net/pipermail/cisco-nsp/ > _______________________________________________ cisco-nsp mailing list [email protected] https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/
