Adaptive Optimal Global Resource Scheduling for a Cloud-Based Virtualized Resource Pool

This paper proposes to employ linear programming algorithms for global resource scheduling to reduce the extra cost for power consumption and operation expenditures, for remote resource access in a cloud-based resource pool with concrete restraints of the networking environment. The scheduler adapts the problem modeling granularity and solution which corresponds to the differential demands of the various stages of a continual process for the initial construction and subsequent operation of a cloud-based resource pool. In particular, the proposed algorithms takes into account resource configuration, service deployment and real-time load, among other factors, to strike a tradeoff among the scheduling performance, response time and computation cost. Different environment modeling methods are provided according to the specific location of networking resource bottleneck. A simple greedy algorithm is provided for a small-scale pool with abundant networking resources.