Modeling of concurrent web sessions with bounded inconsistency in shared data

Client interactions with web-accessible network services are organized into sessions involving requests that read and write shared application data. When executed concurrently, web sessions may invalidate each other's data. Allowing the session with invalid data to progress might lead to financial penalties for the service provider, while blocking the session's progress will result in user dissatisfaction. A compromise would be to tolerate some bounded data inconsistency, which would allow most of the sessions to progress, while limiting the potential financial loss incurred by the service. This paper develops analytical models of concurrent web sessions with bounded inconsistency in shared data, which enable quantitative reasoning about these tradeoffs. We illustrate our models using the sample buyer scenario from the TPC-W benchmark, and validate them by showing their close correspondence to measured results in both a simulated and a real web server environments. We augment our web server with a profiling and automated decision making infrastructure which is shown to successfully choose the best concurrency control algorithm in real time in response to changing service usage patterns.