We introduced a new computation model that reflects the assumptions of the map-reduce framework, but allows for networks of processes other than Map feeding Reduce. We illustrated the benefit of our model by showing how to improve the computation of the 3- way join and by developing algorithms for merging and sorting. The cost measures by which we evaluate the algorithms are the communication among processes and the processing time, both total over all processes and elapsed (i.e., exploiting parallelism). The input file is hold on as a group of files on persistent storage. The goal of a sorting system is to rework this input file into associate ordered set of output files, conjointly hold on persistent storage, such the concatenation of those output files so as constitutes the sorted version of the input file. Our goal is to style and implement a sorting system that may sort datasets of the targeted size whereas achieving a good exchange between speed, resource utilization, and cost.

Author Name: Ms. Neha Chaturvedi, Mrs. Richa Jain ,Mr. Anil Pimpalapure

URL: View PDF

Keywords: Big Data, Cloud Computing, Hadoop, Mapreduce, deep sort method

ISSN: 2455-6203

EISSN: 2455-6203

EOI/DOI:

Add Citation Views: 1

---