Computing on Large Scale Distributed Systems

Since the late 1990's, Desktop Grid systems, such as SETI@Home, have been the largest and most powerful distributed computing systems in the world, offering an abundance of computing power at a fraction of the cost of dedicated, custom-built supercomputers. Many applications from a wide range of scientific domains - including computational biology, climate prediction, particle physics, and astronomy - have utilized the computing power offered by Desktop Grid systems. DG systems have allowed these applications to execute at a huge scale, often resulting in ma jor scientific discoveries that would otherwise had not been possible.

The computing resources that power DG are shared with the owners of the machines. Because the resources are volunteered, utmost care is taken to ensure that the DG tasks do not obstruct the activities of each machine's owner; a DG task is suspended or terminated whenever the machine is in use by another person. As a result, DG resources are volatile in the sense that any number of factors can cause the task of a DG application to not complete. These factors include mouse or keyboard activity, the execution of other user applications, machine reboots, or hardware failures. Moreover, DG resources are heterogeneous in the sense that they differ in operating systems, CPU speeds, network bandwidth, memory and disk sizes. Consequently, the design of systems and applications that utilize these system is challenging.

XtremWeb is somewhere in between pure Desktop Grid system, a la Condor and pure Volunteer Computing system, a la BOINC.

Compared to Condor, we assume that computing resources can be spread over the Internet. XtremWeb relies on a pull model (workers pull tasks from the server) where Condor relies on a push model (the Matchmaker selects the best machine to run the job and push the job on it). XtremWeb doesn't rely on a shared File system and is easier to setup than Condor.

Compared to BOINC, we assume that each user or node has tha ability (if authorized) to submit job. With BOINC, only the project are able to create tasks and insert new application. As a consequence, only few applications are supported by BOINC. In contrast, XtremWeb offers to scientific users an interface similar to batch system (PBS, Condor etc..) to create their jobs and get their results. Also application do not require to be modified. On the other hand, XtremWeb oes not offer credits base system, nor user forum, which can be found elsewhere.