I was wondering, since I do not have experience with blade-based servers, if all the hardware nodes on a blade center act as a single system? I guess they donŽt, since all the blades seem to act like separate servers, requiring software based clustering in order to have them act as fail-over devices as an example, however I read somewhere that latest bladecenters allow an hardware based clustering. Anyone ever heard or had experience with this?
A blade chassis does not form a single system unless you run the right software on the blades (and usually upgrade the networking to Infiniband). You will see blade-based systems in the top 100 but this is through running cluster software rather than anything special about the hardware aside from density, blades are individual servers.
Thank you very much, you have all been very helpfull.
The idea would be to implement a Virtual Environment with failover and scability capabilities and without any needs of special software clustering configuration, in order to make the OS behave completely transparently to the Virtual Server Hypervisor.
Spudstr, thank you, IŽll take a look at those models.