Le mercredi 9 juin 2010 19:01:32, Amos Jeffries a écrit : > On Wed, 9 Jun 2010 14:46:44 -0500, Luis Daniel Lucio Quiroz > > <luis.daniel.lucio@xxxxxxxxx> wrote: > > Le mercredi 7 avril 2010 22:59:59, Amos Jeffries a écrit : > >> On Thu, 08 Apr 2010 08:49:07 +0530, senthilkumaar2021 > >> > >> <senthilkumaar2021@xxxxxxxxx> wrote: > >> > Hi > >> > > >> > What is the use of multicast option in squid ? > >> > > >> > How to configure squid to multicast messages and receive multicast > >> > messages? > >> > > >> > In order to configure multicast whether any configuration is needed > > in > > >> > other than squid > >> > >> http://wiki.squid-cache.org/Features/MultiCast > >> > >> Amos > > > > Amos, i have a doubt about this > > > > in my lan 192.168.1.0/24 we are planning to set up squids in > > workstations > > > so > > we are planning to use multicasting so if caches may be shared. > > > > as we don tknow what ips are connected or not > > it is possible to stay a line like this? > > cache_peer 192.168.1.0 sibling 3128 3130 ttl=5 > > > > or how can we manage this without putting all each ip in all conffs > > I have not dealt much with multicat myself. AFAIK, at present you can't. > But I can't see any real reasons why not. > > The multicast reduces bandwidth for ICP requests going out, but once a > reply comes in it still needs to be linked to a sibling peer. > > At this point some experimenting (or advice from Henrik maybe) is needed > to follow it further. The linking I think looks up an explicitly configured > peer based on the replying source IP. It may be useful to make the peer > dynamic if sufficient security protection can be added. > > Amos So i understand by now it is kind imposible to do a cache_peer 10.200.0.0 sibling 3128 3130 multicast-responder nor cache_peer 10.200.255.255 sibling 3128 3130 multicast-responder