A game theoretic formulation of the technical debt management problem in cloud systems

Abstract

The lack of models to aid the technical debt estimation and monitoring in cloud architectures motivates the need to develop a technical debt management (TDM) theory and practical approach for effective management decision making. The cloud is a marketplace in which services and resources are leased off based on the volume of data and the number of cloud-oriented users, which affect the accuracy of the technical debt measurement. In this research paper, we propose a game theoretic formulation of the technical debt management problem on cloud-based service level. A technical debt measurement game is constructed, parameterizing the current number of players per service, while each new end-user can choose any of the offered cloud-based services. In this direction, a technical debt-inspired throttling mechanism is also discussed for resource leasing optimization and scalability. The validation tests verify the effectiveness of our approach in real-world operations, avoiding accumulated technical debt and optimizing available resources in cloud-centric systems.