An in-browser implementation of Google's MapReduce framework. Using node.js and redis.
JavaScript Python Shell
Latest commit 246b6bc Apr 8, 2012 @spoletto Adding README.
Permalink
Failed to load latest commit information.
client
distributed_testing
results
server_deployment
static
testing
.gitignore
Arke.pdf
README.md
git-push-all.py
index.html
logging.js
redis_db.js
server.js
upload.html

README.md

Arke: An in-browser implementation of MapReduce

Arke is a JavaScript implementation of Google’s MapReduce. Rather than running on a fixed cluster in a data center like most distributed computational frameworks, Arke relies on the computing power of individual browsers across the open Internet. Volunteers connect to Arke’s web application, offering their web browser engine as a worker in the system. Further, Arke is a service; users may upload jobs and data conforming to a simple specification in order to delegate their computation to a dynamic cloud of volunteer nodes.

Arke was implemented by Peter Wilmot, James Sedgwick and Stephen Poletto in the fall of 2011. A report detailing Arke's infrastructure and performance can be downloaded here.