A Dataset and a Comparison of Out-of-Order Event Compensation Algorithms

Event processing is order and time sensitive and therefore assumes temporally correct ordered event streams, even in distributed systems, to be able to create correct results. In this work we discuss implementations of four different out-of-order event compensation algorithms that use different kinds of dynamic time-out buffering techniques, and we compare those to a static buffering method. This is an approach which is generally applicable and easy to integrate for existing distributed systems such as for Internet of Things applications. For the evaluation, specific datasets were recorded, which are introduced in this paper, and which are freely available under a Creative Commons license. Results show, that even with a restrictive buffer setting, most of the out-of-order events can be compensated. Dynamic time-out buffering is still a trade-off between reaction time and out-of-order event compensation, but it is useful in various applications.