Indexing moving objects usually involves a great amount of updates, caused by
objects reporting their current position. In order to keep the present and
past positions of the objects in secondary memory, each update introduces an
I/O and this process is sometimes creating a bottleneck. In this paper we
deal with the problem of minimizing the number of I/Os in such a way that
queries concerning the present and past positions of the objects can be
answered efficiently. In particular we propose two new approaches that
achieve an asymptotically optimal number of I/Os for performing the necessary
updates. The approaches are based on the assumption that the primary memory
suffices for storing the current positions of the objects.
An asymptotically optimal real-time locking protocol for clustered scheduling under suspension-aware analysis