Resource scheduling has been one of the key challenges facing both academia and industry ever since the inauguration of cloud computing. Most of the existing research and practices have been focused on the maximization of the profits of cloud providers, whereas attention to the real needs of cloud users has largely been neglected. In this research, we propose a resource scheduling mechanism empowered with a relevance feedback network, which can be employed by a cloud provider to better meet a user’s resource needs. Our approach is a continuous refinement process that involves three stages: resource matching, resource selection, and feedback integration, where the feedback integration stage allows the resource scheduling history of a user to be considered to update the user’s resource demands and preference. The feedback information integrated in one cycle will effectively adjust the resource matching and selection in the next cycle. Incrementally, this mechanism will produce resource selections that are closer and closer to the user’s real needs. Simulation results indicated that this relevance feedback scheduling mechanism is very effective in satisfying users’ diverse requirements, and it also performs well in terms of the resource utilization rate from the cloud provider’s perspective.
All Science Journal Classification (ASJC) codes
- Theoretical Computer Science
- Information Systems
- Hardware and Architecture