Resource Management in Volunteer Computing Grids

The success of Volunteer Computing (VC) grid applications depends greatly on the systems ability to effectively manage the resources available to it. The most popular middleware application, Berkeley Open Infrastructure for Network Computing (BOINC), has achieved its popularity in part by succeeding in this task. The defining characteristic of VC grids is node unreliability. Because nodes can appear and disappear at any time, it is likely that assigned tasks will not be completed. Various VC architectures have been implemented to deal with this issue. The architecture implemented by BOINC is to send each task to some number of worker nodes; this can be called the redundancy coefficient.