Challenge Overview
At a high level, our goal for clustering is to provide horizontal scalability. This means that we can add more nodes to a system to cope with poor performance instead of adding more physical RAM or CPUs to an existing system (Vertical Scaling). A cluster also provides the benefit of failover. This means that if a node in a cluster were to fail the other nodes in the cluster would pick up the traffic from the failed node without stopping the system. Lastly a cluster also guarantees better uptime for maintenance because you can take out a node in a cluster to upgrade/fix it, while leaving the other nodes untouched. As a result, there will be close to zero downtime for an application using clustering.
However, due to network constraints and lack of container level expertise in our organization we have issues achieving the goals above.
The purpose of this contest is to generate several great architecture ideas, based on the documentation provided, and on dialog with the Kratos stakeholders.
Final Submission Guidelines
Please refer to the wiki spec for the full requirements:
http://apps.topcoder.com/wiki/display/docs/Kratos+Architecture+Idea