Measuring and Diffusing Data Quality in a Peer-to-Peer Architecture

Data quality is becoming an increasingly important issue in environments characterized by extensive data replication. Among such environments, this article focuses on cooperative information systems (CISs), for which it is very important to declare and access quality of data. The article describes a general methodology for evaluating quality of data, and the design of an architectural component, named quality factory, that implements quality evaluation of XML data. The detailed design and implementation of a further service, named data quality broker, are presented. The data quality broker accesses data and related quality distributed in the CIS and improves quality of data by comparing different copies present in the system. The data quality broker has been implemented as a peer-to-peer service and a set of experiments on real data show its effectiveness and performance behavior.