V. Conclusion and Future Work
Credit Union Cloud Model, which aims at tapping into the underutilized computing resources available within an organization/community rather than dedicated servers, provides a promising alternative Cloud Computing solution for organizations and communities. Our work demonstrates that the “no data center” solution indeed works. Besides proving the concept, model, and philosophy of CUCM, our experimental study turned out to be highly encouraging – the “no data center” solution can gain highly competitive performance compared to its counterpart that depends on dedicated cloud servers.
The most significant aspect of our work so far is that by cuCloud we have setup a platform and have a door widely open for many exciting new research issues for the future. The resource pool over which VMs run in cuCloud is not dedicated but shared with native users/tasks. We need to devise a mechanism to provide cloud services reliably and efficiently, while keeping the services from interfering with the native users/tasks at member nodes. Another requirement of CUCM is to have a robust, dynamic and efficient resource management and provisioning mechanism. The resource management and provisioning module should consider the dynamic and unreliable nature of the member hosts that contribute resources to the resource pool of cuCloud. We also need to investigate novel and efficient scheduling algorithms that consider the availability, location, and reliability of the member nodes used by the system to deploy VMs. In addition, cuCloud requires strong security measures in order to insure the security of member nodes from malicious cloud client processes and client VMs from malicious native users at member nodes.
