SCRUTINIZING MAPREDUCE MECHANISM WITH ORIENTATION TO REFINE THE PRODUCTIVITY

Rakshitha.N, Prasanna Kumar.M

Abstract: In Today’s fast evolving world, number of users of internet is increasing at speed at which light travels. This is directly proportional to requirement of storage. Administering this wide range of data which is commonly called as big data is laborious. Big data is bevy of both structured and unstructured data. Dealing with this big data broach new challenges. Map reduce is the key to tackle these strenuous situations. Map reduce is the most favored evaluating methodology for colossal data processing in disseminated milieu. It is admired and widely accepted because of its outstanding features. Map reduce is hardcore in hadoop. Hadoop is the open source implementation of map reduce framework. It is used for distributed storage and processing of very large data sets on computer clusters built from commodity hardware. Map reduce is combination of map and reduce functions. Mapping and reducing the task based on slot is accomplished to improve momentum. Maps reduce fails to exploit all advantages because of unoptimized resource allocation and utilization. To unravel this we proffer to allow slots to be vigorously awarded to either map or reduce tasks depending on actual need. These designation of slots accomplished break the traditional rules and gave new dimension to overcome the loop holes of slot based methodology. This affects positively and aids to uplift performance of map reduce performance.

Keywords: Big data, Map reduce, Resource allocation.

DOI: https://doi.org/10.15623/ijret.2015.0403074