An international group of researchers has developed Misco, a version of Google's MapReduce algorithm powered by cell phones in a self-contained cloud computing environment. In a test, about 20 smartphones were employed to create a server farm for Misco, which enabled the distributed processing of large amounts of data to take place closer to the data itself. With Misco, the data does not have to be transmitted first over a relatively slow wireless network.
In some instances, the system can process the data even faster than if it is first uploaded, in total, to a remote server, which would be much faster than the processor on any one phone.
MapReduce is difficult to implement on cell phones partly because in a regular server farm the failure rate is relatively low, and the latency of a signal transmitted between servers also is low. As a result, the speed of the Misco processing network would degrade exponentially as the failure rate of any one node in the system increases.
From Technology Review
View Full Article
Abstracts Copyright © 2010 Information Inc., Bethesda, Maryland, USA
No entries found