IBM White Paper Sample

This paper presents and evaluates Blue Eyes, a scalable and reliable system management solution based on a multi-server architecture. We described how multiple management servers are organized effectively and reliably manage ever increasing endpoint systems requiring administration.

Management workload is evenly distributed by balancing the number of endpoints covered by each management server. Management tasks are forwarded efficiently to endpoints through a tree structure (MST) of management servers. Secondary servers organized in a logical ring (MRR) help the system be reliable and highly available.

We implemented a prototype of Blue Eyes with a core set of server-server operations and distributed mechanisms to support a basic management tasks. It is important to note that this implementation does not impose significant modification of management task code of the legacy management solutions.

Our early experimental results, using the prototype, show that Blue Eyes can be expanded with balanced workload for endpoints, good performance characteristic for management tasks and a small scale-out overhead.

We also assessed how the system resource utilization changes over a fail-over process.