A priority-based transmission protocol for congested networks supporting incremental computations

Abstract

Many fields of human activity produce data at an increasing rate, that have to be analysed in an effective and timely way, both locally at the data sources and globally, where decisions have to be taken. Very often in distributed applications, delays originate from the limited network bandwidth or temporary network congestions. To cope with these problems and to satisfy the requirements of online interactive applications (timely availability of data whose quality could eventually be reduced), new compression algorithms and transmission protocols that can trade off the exactness of results against their timeliness, have to be developed. This paper presents a priority-based data transmission protocol for sending large volumes of data (data cubes) over congested networks, that allows the incremental implementation of OLAP applications. A wavelet-based lossy compression algorithm that is the core of the transmission protocol is also presented.