Pando: Personal Volunteer Computing in Browsers

The large penetration and continued growth in ownership of personal electronic devices represents a freely available and largely untapped source of computing power. Moreover, the large environmental and social impact of producing these devices suggests we should better use those that already exist. We aim to make these devices available for parallel computations to both scientists and other programmers of the general public, for their personal projects, and in the simplest way possible to program and to deploy. We named our approach to distributed computing personal volunteer computing. We designed, implemented, and tested Pando, a new distributed computing tool based on a declarative concurrent programming model, organized around the pull-stream design pattern, and implemented using JavaScript, WebRTC, and WebSockets. This tool enables a dynamically varying number of failure-prone personal devices contributed by volunteers to parallelize the application of a function on a stream of values, by using the devices' browsers. To illustrate Pando's capabilities, to show its benefits as well as its limitations, we implemented a variety of applications including crypto-currency mining, hyper-parameter optimization in machine learning, crowd computing, and open data processing and tested it using diverse devices we have accumulated over the years. Pando, both as a tool and a reference design, should therefore be a useful addition to the parallel toolbox of a multitude of users and a complementary approach to existing parallel and distributed computing alternatives.