Performance Issues

Abstract

This chapter is concerned with performance issues, the management of which may become complex with respect to large-scale systems and particularly those which incorporate biometric identity verification. In order to understand such issues, we must define the scope of our performance definition, the related metrics and consider how we are to monitor performance overall. This chapter discusses such issues, even from the point of biometric capture, with its related user and technical factors, through biometric operation, including user factors, technical factors, the biometric matching process, common metrics and an explanation of the biometric threshold setting. In this context, the User Psychology Index (UPI) is introduced and examples are given of the distinction between theoretical performance and realised performance. A new metric is also introduced in the form of the average error rate (AER), a single figure which provides for a simple comparison of device or configuration scenario realised performance.

Database performance is discussed in some detail, including factors such as the use of multiple databases, location, network considerations, configuration, database ­optimisation, query optimisation, read–write performance and other relevant factors. Leading naturally from this topic is network performance, with factors such as bandwidth and latency, associated configuration, network monitoring and the use of third party networks. We then move on to operational performance and discuss functional performance from the human perspective, including the effect upon transaction times and systems throughput. We additionally consider performance monitoring and documentation, including centralised log management for system events and the biometric matching score for biometric identity verification transactions. We introduce the concept of user psychology, as pioneered by the author, and discuss user variables including usability, familiarity, competence, attitude, external influence, disabilities, language, age, ethnicity and well-being, all of which may have a measurable effect upon realised performance. Lastly, we consider the finer points of systems configuration and introduce the concept of total systems performance. This chapter provides a good grounding in the performance of systems which incorporate ­biometric functionality, from a very practical perspective.