Enhanced genetic algorithm to reducemakespan of multiple jobs in map-reduceapplication on serverless platform
Divya Thorat · NORMA · 2020
Nowadays, allocating proper tasks to the resources is an integral part of the cloud environment. So the execution time is depending on the number of resources allocated to the environment. So it necessary to choose the proper scheduling algorithm for multiple applications. The serverless platform is the combination of function as a service and back-end as service. This paper proposed a map-reduce jobs with a genetic algorithm on a serverless platform. In this, we have used Lambda function as a service and s3 bucket, Redis storage as back end as service. The combination of fast and slow storage gives the fine-grained elasticity, and a genetic algorithm minimizes the total execution time. In a serverless platform, The pricing depends upon the number of times the function executed and the total execution time required for operation. The genetic algorithm requires low execution time, so it reduces the cost of the operation, and a combination of slow-fast storage gives better performance along with efficiency. The evaluation carried out on a serverless platform, comparison carried out between map-reduce application without genetic algorithm and with a genetic algorithm, so result shows the map-reduce application with a genetic algorithm requires less execution time and has higher performance.