GitHub - gsastry/hadoop-mapreduce: Mirror of Apache Hadoop MapReduce

Branches Tags

Name		Name	Last commit message	Last commit date
Latest commit History 867 Commits
.eclipse.templates		.eclipse.templates
.externalToolBuilders		.externalToolBuilders
bin		bin
conf		conf
ivy		ivy
lib/jdiff		lib/jdiff
src		src
.gitignore		.gitignore
CHANGES.txt		CHANGES.txt
LICENSE.txt		LICENSE.txt
NOTICE.txt		NOTICE.txt
README.txt		README.txt
build-utils.xml		build-utils.xml
build.xml		build.xml
ivy.xml		ivy.xml

Repository files navigation

FlowScheduler: A New Task Scheduling Algorithm for Hadoop/MapReduce
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

by: Girish Sastry

FlowScheduler is a new task scheduler for hadoop, optimal within an
additive constant. It consists primarily of two phases:

1) max-cover: a flow augmenting algorithm based of Fold-Fulkerson
2) bal-assign: a greedy task assignment algorithm that exploits data locality
    and network latency