Personal Volunteer Computing
share
Since the 1990s, the number of personal computing devices has exploded. Today we collectively own billions of devices whose computing resources are underused the majority of the time. These resources are either time-consuming to leverage or not supported by other mainstream distributed computing approaches. We propose to revisit volunteer computing, where volunteers from around the world contribute the computing resources of their devices for high-profile projects, with a more personal focus along these main dimensions: nature, scope, and length of the project tasks; relationship with the volunteers; ownership of the participating devices; and ownership and complexity of the platform and technologies. We present our position in contrast to other popular distributed computing approaches, in order of increasing similarity. Our approach is both original and it highlights the historical lack of focus on the limiting effects of complexity on the computing platforms: this complexity limits the ability of their users to deploy and repurpose the platforms in evolving contexts. To show the viability of the approach we implemented Pando, a personal volunteer computing platform for the Web. This tool now enables scientists and programmers to quickly use their personal devices and those of their friends for accelerating their computations by writing a single JavaScript function. It can be described with only a few abstractions. Tools built for personal volunteer computing, such as Pando, may reduce the need for acquiring additional devices, and therefore reduce pressure on resources needed to produce them, because they tap in the abundance of computing resources that already exist. Moreover, they may spread the benefits of computing more evenly by serving the needs of communities and individuals that are not served by other mainstream approaches.
READ FULL TEXT