Sandia National Laboratories
We have been using the JPPF since December 2007, and we have found the framework to be extremely powerful and easy to work with. The JPPF distributed framework has enabled us to create and maintain a network of over 140 processors for use in computationally intensive applications. Currently we use the framework to calculate predicted seismic wave travel times through global 3-D Earth models. As with any tool as complex as a distributed computing package (JPPF), understanding the multitude of parameter setting options can be challenging. The very responsive JPPF support team and user forums have been a crucial resource for answering our system related configuration issues. JPPF is a very powerful package!
Sandia National Laboratories
We use JPPF on small clusters with up to 20 nodes to handle computationally expensive, non-recurring tasks that include statistical analysis, basic linguistic analysis, genetic algorithms for neural networks and classifier training. Not only is JPPF easy to setup and use, but the task-oriented design is very powerful yet approachable and easy to understand. This ease of use is JPPF's greatest strength because it effectively allows us to utilize our existing code, processes and algorithms in a distributed environment with minimal effort and time. The ability to adapt our existing technology without having to redesign or rethink entire processes is fantastic. In addition, the forums have provided incredibly fast and useful assistance when necessary. Overall, we have found JPPF to be a very practical and valuable tool in our toolbox.
During my work I used the JPPF framework to create a distributed computing environment for the existing Eclipse RCP application. To distribute existing code I always went through the process of isolating the task from other code so that it would be able to run on a separate environment. JPPF works in a master/slave principle using the concept of clients, drivers and nodes. To gain optimal performance and solve class loading issues the entire framework was embedded in the [...] OSGi runtime. [...] JPPF is a solid framework built for transparent distributed computing in Java. It can provide many topologies through simple settings in the configuration file. [...] JPPF is very mature in the creation of topologies because of the driver principle. It is a framework that is designed to create topologies.
JPPF has been the main framework used for the development of Distributed Java Hash Cracker project, a program designed to make brute force attacks. DJHACK carries out intensive mathematical calculations. Thanks to the ease of use of JPPF, it has been really easy to develop a distributed environment for this application. Through scalability, granularity in the configuration, and traffic monitoring tools provided by this framework, a robust and complete application has been developed. But this would not have been possible without the collaboration of JPPF developers, who have been able, through the forum, to resolve all doubts and questions that have been sent during the application development.
For all this, I can only say that JPPF is the framework for excellence in parallel computing.
Unai Gómez Velasco
|Copyright © 2005-2013 JPPF.org|