Comparative Study of Map Reduce Task Scheduling Optimization Techniques
Vishal Kumar, Sumit Kushwaha · 2023
As data processing capabilities improved, the flow of data increased significantly, necessitating the use of distributed systems and a new programming framework, MapReduce, to manage this massive amount of information. Any job requiring a large amount of data must be divided into subtasks for parallel execution. In order for all subtasks to be executed concurrently need to be scheduled parallelly on available set of virtual machines. This is not an easy task; it requires consideration of a number of critical parameters such as make-span, Execution cost, risk of execution, energy consumption and so on. The optimization of these parameters is critical for effective scheduling. The most effective hybrid scheduling techniques have been developed and implemented on a single platform with a diverse set of virtual machines. On a single platform, it compares all well-known and popular scheduling techniques. All experimental values for various parameters of interest for all best-ever scheduling techniques will be recorded for analysis, and results representing parameter-wise best-available scheduling techniques will be presented. This comparative study aids in the direction of future research and serves as a guide for progress from the current state of the research.