On Mon, 1 Sep 2003, Michel Arboi wrote: > Maybe 192.0.34.166 (= example.com) which BTW is _not_ part of test-net > (192.0.2.0) as the RFC said. or maybe an address in test-net? Hm, neither of them should be used by anyone (for anything important), so you can test either of them...or even both of them. > > Perhaps the infamous "link local" 169.254.0.0/16? > What's the use of this, exactly? I understood that it is a backup > address pool when no address has been set and no DHCP answers. A host can pick an address from this block and use it to talk to other hosts on the same LAN (that have picked their own address of this kind). The result is an "autoconfigured" isolated IP network. MS Windows boxes default to this (truly redmondian) mode of operation when they fail to obtain a DHCP lease. A nice property of these addresses (in the context of RIP testing) is that they are not supposed to be routable. On the other hand, the same property might cause false negatives and example.com or a similar block of "more innocent-looking addresses" might be a better choice. On Mon, 1 Sep 2003, Michel Arboi wrote: > What about this: > - I set a route to example.com > - I remove it after the test. The bogus route should time out in a few minutes (the router has a BIG problem otherwise). I am not sure explicit removal will make it disappear considerably sooner (esp. in the case when the route would be propagated to other routers and esp). --Pavel Kankovsky aka Peak [ Boycott Microsoft--http://www.vcnet.com/bms ] "Resistance is futile. Open your source code and prepare for assimilation."
This archive was generated by hypermail 2b30 : Mon Sep 01 2003 - 15:04:31 PDT