A Web Computing Environment for Parallel Algorithms in Java

Abstract

We present a web computing library (PUBWCL) in Java that allows to execute tightly coupled, massively parallel algorithms in the bulk-synchronous (BSP) style on PCs distributed over the internet whose owners are willing to donate their unused computation power. PUBWCL is realized as a peer-to-peer system and features migration and restoration of BSP processes executed on it. The use of Java guarantees a high level of security and makes PUBWCL platform-independent. In order to estimate the loss of efficiency inherent in such a Java-based system, we have compared it to our C-based PUB-Library.

As the unused computation power of the participating PCs is unpredictable, we need novel strategies for load balancing that have no access to future changes of the computation power available for the application. We develop, analyze, and compare different load balancing strategies for PUBWCL. In order to handle the influence of the fluctuating available computation power, we classify the external work load.

During our evaluation of the load balancing algorithms we simulated the external work load in order to have repeatable testing conditions. With the best performing load balancing strategy we could save 39% of the execution time on average and even up to 50% in particular cases, in our test environment