I'm very interesting in purchase an pair of servers to create a H.A. cluster.

The minimal scheme that I would like to implement is one server running the services and another waiting to enter in service when the running fail ( hearbeat... )....

Basically what I need is a public IP that wil be in use for the running server, if that server fails the witing server will setup that public IP to recive the incoming traffic and services.

i'm wondering if the network architecture of the data center would alllow me to change that ip and recive the traffic of the failed server.

Other scheme would be a load balancing servers capable to use the scheme of getting the service IP ( a private in that case ) if detects that other computers are down. This scheme would schew that the traffic for a fallen server would be catched for other servers. The network requeriments are the same of the last scheme.

Could somebody help ?

Thank you very much in advance.

Wouldnt be the second scheme exactly what EV1 offers?

http://www.ev1servers.net/english/high_ava...il_clusters.asp

I would need a confirmation of that, as I understand that the "balancer" couldn't to be enough smart to detect a failure of the computers.

If the balancer is not smart, I would need to switch the IPs automatically so the other computers can get the traffic sent by the balancer to the failed server. ( the schemes porpoused in my last post )

I'm sorry but I can't find documentation related to my questions in the page you posted me, I just find prices.

Thank you very much for you reply.

Look at this post:

http://forum.ev1servers.net/showthread.php?t=48145

Thanks again.
Bye.