Monadic Logs for Collaborative Web Applications

Cloud based web-applications are quickly becoming common in modern society. A new class of such applications, collaborative cloud applications, are gaining in popularity as they greatly improve remote collaboration. Most of these applications use a log structure as a coordination mechanism for shared application state. Such structures typically store the entire application state as well as deltas (changes sets) while the application runs. In this paper we propose a monadic, dependency-aware, self-cleaning log structure for collaborative cloud applications, which we refer to as a monadic log. This structure provides a rich set of analytical tools to support a variety of log transformations and rewrites. For example, the garbage collection mechanisms already present in any managed language will automatically bound the memory footprint of a monadic log. Moreover, a monadic log substantially eases the computational and bandwidth burdens of a server infrastructure when compared with traditional log structures.