Decentralized self-adaptation in large-scale distributed systems

Abstract

The evolution of technology is leading to a world where computational systems are made of a huge number of components spread over a logical network: these components operate in a highly dynamic and unpredictable environment, joining or leaving the system and creating connections between them at runtime. This scenario poses new challenges to software engineers that have to design and implement such complex systems. We want to address this problem, designing and developing an infrastructure, GRU, that uses self-adaptive decentralized techniques to manage large-scale distributed systems. GRU will help developers to focus on the functional part of their application instead of the needed self-adaptive infrastructure. We aim to evaluate our project with concrete case studies, providing evidence on the validity of our approach, and with the feedback provided by developers that will test our system. We believe this approach can contribute to fill the gap between the theoretical study of self-adaptive systems and their application in a production context.