OurGrid
| Company type | Open-source software project |
|---|---|
| Industry | Grid computing, Middleware |
| Founded | December 2004 |
| Headquarters | Campina Grande , Brazil |
| Products | Grid middleware software |
| Website | www |
OurGrid represents an open-source grid middleware solution constructed on a peer-to-peer framework, with its primary development taking place at the Federal University of Campina Grande in Brazil. Since December 2004, the university has maintained an OurGrid instance, enabling users to harness significant computational capabilities by pooling their idle resources. Participants who contribute a larger share of resources gain increased access when they require it. This platform facilitates the execution of parallel applications composed of independent tasks, making it ideal for simulations, data mining, and search-related activities.
Overview
Implemented in Java, OurGrid allows users from any operating system compatible with the Java virtual machine to join the network. Its architecture consists of four main elements: Broker, Worker, Peer, and Discovery Service. The Broker is responsible for managing user computations; the Worker provides unused resources to enhance network reputation; the Peer oversees multiple Workers; and the Discovery Service supports resource sharing among Peers.
To safeguard against malicious actions such as DDoS attacks, the Worker integrates virtualization technology that isolates running tasks from both the host system and the Internet.
Incentive mechanism
OurGrid utilizes a Network of Favors strategy to discourage users from merely consuming computational power without reciprocating. By contributing idle processing capacity, users build a reputation that prioritizes their computation requests. This reputation is maintained locally by peers who have engaged with the user and is designed never to drop below zero, preventing abuse through identity changes.
Limitations
Research conducted in 2013 by Marek Šimon, Ladislav Huraj, and Vladimír Siládi highlighted certain performance constraints within P2P grid systems like OurGrid. Tasks involving substantial data distribution overhead—such as snow cover depth interpolation—did not experience efficiency improvements compared to sequential execution. In contrast, compute-heavy tasks with minimal data transfer demands—like simulations of radioactive decay—showed enhanced performance when parallelized using the grid.