A Framework for Parallelizing Large-Scale, DOM-Interacting Web Experiments

Concurrently executing arbitrary JavaScript on many webpages is hard. For free-standing JavaScript — JavaScript that does not interact with a DOM — parallelization is easy. Many of the parallelization approaches that have been applied to other mainstream languages have been applied also to JavaScript. Existing testing frameworks such as Selenium Grid [19] and PhantomJS [16] allow developers to parallelize JavaScript tests that do interact with a DOM. However, these tools frequently limit the user to very constrained programming models. Although it is possible to build testing programs that take arbitrary inputs and yield arbitrary outputs (rather than pass or not pass) on top of these systems, the work required to do so is non-trivial. Controlling inputs — that is, pages — in order to carry out web research on real world pages is even less supported in modern tools. In fact, no existing tool provides same page guarantees. In response to this paucity of flexible and reliable systems, we built a framework designed for running large-scale JavaScript experiments on real webpages. We evaluate the scalability and robustness of our framework. Further, we present the node addressing problem. We evaluate a new node addressing algorithm against preexisting approaches, leveraging our framework’s clean programming model to streamline this large-scale experiment.