Measurement of software operability and its main determining factors

OData support
Supervisor:
Kocsis Imre
Department of Measurement and Information Systems

The IT processes concerning management and operation have a number of well defined approaches for qualification methodologies. Less effort has been invested however in the qualification of the inherent operability of a software system. Furthermore, the factors that influence the inherent operability of a software system are to this day not identified. That is partly due to the context sensitive nature of the definition and of the conceptualization of operability.

In this work, using literature sources, I overview the current approaches to define and quantify the operability of a software system. The conceptualization of operability is also inspected in an enterprise environment in an effort to determine the factors that are suspected to influence it. Based on methods from the literature study and knowledge gained from an enterprise environment I propose an approach to quantify operability. As logging is considered as an important influencing factor of operability I also propose an approach to assess the logging capabilities of a software system. Finally, using data from surveys that I carried out in the enterprise environment, and historic data of incident management, I observe data relationships that reflect on the operability of the inspected set of software systems. The goal of this examination is to find out, whether any relationship between the metrics of the incident management and the metric of operability exists.

The relationship exploration revealed, that there seem to be a connection between the operability metric I propose, and the subjective operability of the softwares in the observed set. Reservations concerning the validity of the connection are discussed in detail. Furthermore it has been revealed that the logging capabilities, assessed using both a subjective and an objective scale, do indeed have a connection with the relative operability metric. To assess the validity of the discovered relationships however, a much larger set of software systems must be examined. At the same time, these initial results show already, that ‘traditional’ design metrics should also be taken into consideration when assessing the operability of a software system.

Downloads

Please sign in to download the files of this thesis.