Providing resources and services on demand, cost effectively and with high performance have been the need of the hour in the MapReduce cloud environment. While the size of data is growing rapidly due to impact of social media, scientific experiments, file logs created by different departments around the globe, etc., industry have started using this data for analysis and orient their work accordingly. Since the services are provided through virtualized resources, the placement of Virtual Machines (VMs) in cloud environment has a major impact on the performance of the service. Recently there have been a number of research activities on Virtual Machine Placement. Models based on Queuing Theory Approach, Matrix Manipulation Approach, Integer Linear Programming Approach, Min Max Approach, Flow-Network Based Approach, and Tree based Approach have been proposed. The major objective of all these schemes is on how to consolidate the VMs on servers to save power and increase performance. This study summarizes the different major approaches used in recent time for Virtual Machine Placement in MapReduce cloud and presents a comparative analysis. This paper also highlights the major contributions made so far and the open research issues existing in this field.
Kumar, S D Madhu
"A comparative evaluation of VM placement techniques in mapreduce cloud,"
Manipal Journal of Science and Technology: Vol. 1:
1, Article 3.
Available at: https://impressions.manipal.edu/mjst/vol1/iss1/3