Vanishing Point: Where Infrastructures, Architectures, and Processes of Software Engineering Meet

In software project management, there exists a triangle-like relation connecting the required time to deliver a certain scope of software features with a certain cost. If one of these three is affected, others must compensate for this. For example, having a faster delivery means either a costlier product or less features, or both. Long delivery times are usually unacceptable in any case as the business environment is changing fast.

To deal with this, contemporary software is mostly produced with Agile methods, which emphasise developing small increments to deliver constant stream of value to the customer. A small piece of software is easier to produce and test. Also, rapid feedback can be gained with tight co-operation with the customer. As the increments have become almost infinitesimally small, the working software can be constantly improved as the changes can be delivered to the end user almost instantly. However, this is only possible when the increments are reliably tested and the delivery itself is rapid. Thus, automation in these crucial parts is a must. Furthermore, the customer is not able to comment on every change in person, so the collection of the feedback must be automated. Furthermore, the software product itself has to support the continuous delivery. There exists a certain relation between these aspects–namely the tool infrastructure, processes and the architecture—reminiscent of the project triangle of time, cost, and scope.

In this thesis, we examine the crucial properties these aspects in the context of increasing the speed of delivery–up to continuous delivery and deployment combined with the idea of continuous feedback. Also, the ramifications of rapid software delivery are studied. The research is carried out as interviews and related methods, such as surveys, to gain data from the companies involved in software development. Also, some quantitative analysis is used to back up the findings. As a result, a model is introduced based on the research. It can be used to explore the aspects and their interrelationships. We present a set of key enablers of increasing the delivery speed and present a set of side-effects that have to be considered. These can be used as a guideline in the companies which are striving to hasten their delivery pace. Additionally, a comparison of various companies based on their delivery speed is presented.