An Accelerated Map Reduce Mechanism using Network levitated Merge
| Author(s) | : | Bhosale Vinod Datta, Khutal Ajit Sudam, Kurhade Chetan Nivrutti |
| Institution | : | Department of Computer engineering, Jaihind college of engineering, Kuran,pune |
| Published In | : | Vol. 3, Issue 11 — November 2016 |
| Page No. | : | 255-259 |
| Domain | : | Engineering |
| Type | : | Research Paper |
| ISSN (Online) | : | 2348-4470 |
| ISSN (Print) | : | 2348-6406 |
Hadoop technology could be a standard open ASCII text file of the MapReduce programming model forcloud computing. It faces a multiple of problems to attain the most effective performance from the actual systems. Theseembody a publishing barrier that delays the scale back part, cyclic merges, and disk accesses, and therefore the lack ofmovable ness to completely different interconnects. To stay up with the increasing volume of knowledge sets, Hadoopadditionally needs economical I/O capability from the underlying pc systems to method and analyse information. We tendto describe Hadoop-A, associate acceleration framework that optimizes Hadoop with plug-in elements for quickinformation movement, overcoming the prevailing limitations. A unique network-levitated merge algorithmic rule isintroduced to merge information while not repetition and memory access. Additionally, a full pipeline is meant to overlapthe shuffle, merge, and scale back phases. Our experimental results show that Hadoop-A considerably hastensinformation movement in MapReduce and doubles the output of Hadoop. Additionally, Hadoop-A considerably reducesdisk accesses caused by intermediate information.
Bhosale Vinod Datta, Khutal Ajit Sudam, Kurhade Chetan Nivrutti, “An Accelerated Map Reduce Mechanism using Network levitated Merge”, International Journal of Advance Engineering and Research Development (IJAERD), Vol. 3, Issue 11, pp. 255-259, November 2016.
