resource description in a server-cluster

Question

I am searching for a way to describe resources/services on a server-cluster but couldn't find a proper solution yet although I actually think there should be one.

We're running here a bunch of servers, each of them has is responsible for a different tasks and therefor runs different services. For management of those servers, we would like to introduce a (central) monitor which knows about the resources (such as software, data, specific hardware) of each server. So, technically there is a need for some discovery-mechanism to find each server and some functionality to get this information from the server. Also important is to have the interface to have the interface to get this information very generic, so it can be easily extended.

As I assumed that there is a need for something like that in the cloud-computing area, I already stumbled over OCCI. Although it comes close (and might be a little bit too much for my needs) - I don't see any discovery mechanism here. How is the service-description solved in cloud-computing environments?

Apart from that, I also though about implementing an own solution using REST, Upnp or XML-RPC, but still think something like that has to be out there - at least some xml schemes?!. Maybe I am just using the wrong search terms? I would be happy for any hint.

Answer 1

Have you thought of using Zabbix?

It is open-source, heavily customizable and introduces several hardware and software discovery mechanisms through its agent.

Quoting just a few points from their web-page :

• auto-discovery of servers and network devices
• low-level discovery
• distributed monitoring with centralized web administration

And if you can't do it with Zabbix out-of-the-box, there is a high chance you can implement it easily through templates and agent magic.

Answer 2

Take a look at CompatibleOne, which has the service description (as well as brokering) features that you describe.