A scalable web-based real-time information distribution service for industrial applications

In this paper we discuss a web-based content distribution service for wide-area distributed embedded systems to allow remote monitoring of their state using web browsers, servers, and caches. The service allows browsing the real-time state of remote devices, industrial processes, controllers, appliances, agents, and other entities that may publish their state. For a remote monitoring application, it creates the illusion that all monitored entities are available locally in real-time, as opposed to scattered across a wide-area network with unpredictable backbone delays. The main difficulty is to achieve this in a scalable manner without wide-area network support for real-time communication while simultaneously reducing the amount of backbone traffic (which carries process state updates). Another challenge is to achieve this without significant modifications to the existing web infrastructure. We provide timely response and bounded consistency for the monitored data. We have built a prototypical service architecture based on freely available web servers/proxy caches and shown its basic applicability to industrial applications by the experimental results.