From: Robert K. <r....@cr...> - 2006-01-05 14:52:44
|
On Wed, 2006-01-04 at 17:17 -0700, Tristan RHODES wrote: > Question about the future of PacketFence: > It seems likely that as technology progresses, routing will be pushed > farther out from the core to the access layer. Are there any ideas on > how Packetfence can still be deployed passively, without placing a > server in every building? If not, and we end up needing a PacketFence > server in every building, are there any plans to help with the > management of dozens of PacketFence boxes? Though several vendors have been trying their hardest to get everyone to push routing to the edge I'm not really convinced there's much advantage to it. Should it happen through I don't see why you'd need a server per building. Remember all that packetfence requires is to be in the same layer-2 broadcast domain as the machines it is managing. Given that packetfence already supports 802.1q VLANs there's no real reason you can't arrange for a single packetfence system to see multiple buildings. It might get a bit messy depending on your network layout, but it's not impossible. > I realize that the other alternative is to be inline with traffic, and > route traffic through the PacketFence box. If this is the prefered > method, perhaps it would be beneficial to develop an easy way to deploy > a cluster of PacketFence boxes. I think this would be beneficial regardless, and it might not even be too difficult. There are lots of linux clustering solutions available nowadays and I believe MySQL already has built in replication support. The trouble is it's likely to be really hard to do in a distribution independant way. -- Robert Kerr |