FogBrainX

Continuous Reasoning for managing next-gen Cloud-Edge applications in continuity with the CI/CD pipeline

Developing and releasing multiservice applications rely upon a pipeline of automation tools known as Continuous Integration/Continuous Deployment. Among those tools, Continuous Reasoning is exploited by large companies to perform incremental static analyses on their code commits as soon as they are integrated into a shared codebase. Within this context, we extend continuous reasoning towards the continuous QoS- and context-aware management of multiservice applications in Cloud-IoT scenarios. FogBrainX is a novel continuous reasoning methodology that supports runtime decision on service placement by reacting both to changes in the infrastructure and in the application requirements, and capable of suggesting migrations only for services affected by such changes. Experimental results show that our approach brings considerable speed-up in comparison with an exhaustive search employing non-incremental reasoning.

Repository

  • Publications
  •    S. Forti, G. Bisicchia, A. Brogi. Declarative Continuous Reasoning in the Cloud-IoT Continuum. Journal of Logic and Computation, 2022.
    PrototypeDOIRepositoryCite
    Developing and releasing multiservice applications rely upon a pipeline of automation tools known as Continuous Integration/Continuous Deployment. Among those tools, continuous reasoning is exploited by large companies to perform incremental static analyses on their code commits as soon as they are integrated into a shared codebase. In this article, we extend continuous reasoning towards the continuous QoS- and context-aware management of multiservice applications in Cloud-IoT scenarios. We propose a novel continuous reasoning methodology that supports runtime decision on service placement by reacting both to changes in the infrastructure and in the application requirements, and capable of suggesting migrations only for services affected by such changes. The methodology is prototyped in Prolog and assessed through simulations over a realistic use case and over a lifelike motivating scenario at increasing infrastructure sizes. Experimental results show that our approach brings considerable speed-up in comparison with an exhaustive search employing non-incremental reasoning.