Optimal VM Placement Model for Load Balancing in Cloud Data Centers

摘要

In cloud data centers, there are numerous hosts as well as application requests that have been developed. These requests need to allocate resources dynamically which raises the traffic scalability issue. The authors formulate the idea of selecting the most optimal host allocation scheme which can improve the power consumption as well as resource utilization ratio. We propose an Optimal VM Placement for Load Balancing (OPLB) using Maximum Likelihood estimation for parallel and distributed applications. The problem is formulated in a speculative framework based on CPU, Memory and Energy estimations to capture the improvements in throughput and failure rate. The performance evaluation demonstrates that the proposed method achieves significant traffic scalability improvement upto 49.54%, 32.63% and 19.23% over random, sequential and LB-BC virtual machine placement heuristics respectively.