A framework for building scalable web applications for high-resolution cluster-based display walls

Abstract

As technology advances, researchers in the natural sciences collect ever-increasing amounts of data. While computer science research often focuses on effective ways to perform computations on large data sets, the visualization of large data sets can be just as important for achieving new insights. Just as cluster computing enables scalable computation on large data sets, so can cluster-based display walls enable scalable visualization of large data sets.

At the same time, visualization and user interface libraries are being most extensively developed for the web. Examples of such libraries include D3 for visualization and Bootstrap for user interfaces. Such libraries often contain built-in support for scaling down to small displays (i.e. on mobile devices), however, they have no such support for scaling up to cluster-based display walls, which require coordination among multiple display hosts.

This thesis presents a framework, provisionally named Browzawall, for building web applications that scale up to high-resolution cluster-based display wall environments. Browzawall consists of several JavaScript libraries and a WebSocket server. By using Browzawall, application developers can build web applications that scale from mobile and desktop environments up to cluster-based display wall.