Sector and Sphere: Towards Simplified Storage and Processing of Large Scale Distributed Data - Computer Science > Distributed, Parallel, and Cluster ComputingReport as inadecuate




Sector and Sphere: Towards Simplified Storage and Processing of Large Scale Distributed Data - Computer Science > Distributed, Parallel, and Cluster Computing - Download this document for free, or read online. Document in PDF available to download.

Abstract: Cloud computing has demonstrated that processing very large datasets overcommodity clusters can be done simply given the right programming model andinfrastructure. In this paper, we describe the design and implementation of theSector storage cloud and the Sphere compute cloud. In contrast to existingstorage and compute clouds, Sector can manage data not only within a datacenter, but also across geographically distributed data centers. Similarly, theSphere compute cloud supports User Defined Functions UDF over data bothwithin a data center and across data centers. As a special case, MapReducestyle programming can be implemented in Sphere by using a Map UDF followed by aReduce UDF. We describe some experimental studies comparing Sector-Sphere andHadoop using the Terasort Benchmark. In these studies, Sector is about twice asfast as Hadoop. Sector-Sphere is open source.



Author: Yunhong Gu, Robert L Grossman

Source: https://arxiv.org/







Related documents