Slicing Techniques for Temporal Aggregation in Spanning Event Streams

Abstract Slicing is a popular approach to perform aggregation in streaming systems. It allows sharing computation costs among overlapping windows. However those systems are limited to point events. In this paper, we address the temporal aggregate computation issue in streams where events come with a duration, denoted as spanning event streams. After a short review of the new constraints ensued by event lifespan in a temporal sliding-window context, we propose a new structure for dealing with slices in such an environment, and prove that our technique is both correct and effective to deal with such spanning events. We then further extend this technique to compensate for the new constraints induced by the duration of spanning events with a multi-level structure able to reduce insertion costs.