G<sc>o</sc>P<sc>rime</sc>: A Fully Decentralized Middleware for Utility-Aware Service Assembly

Abstract

Modern applications, e.g., for pervasive computing scenarios, are increasingly reliant on systems built from multiple distributed components, which must be suitably composed to meet some specified functional and non-functional requirements. A key challenge is how to efficiently and effectively manage such complex systems. The use of self-management capabilities has been suggested as a possible way to address this challenge. To cope with the scalability and robustness issues of large distributed systems, self-management should ideally be architected in a decentralized way, where the overall system behavior emerges from local decisions and interactions. Within this context, we propose GOPRIME, a fully decentralized middleware solution for the adaptive self-assembly of distributed services. The GOPRIME goal is to build and maintain an assembly of services that, besides functional requirements, fulfils also global quality-of-service and structural requirements. The key aspect of GOPRIME is the use of a gossip protocol to achieve decentralized information dissemination and decision making. To show the validity of our approach, we present results from the experimentation of a prototype implementation of GOPRIME in a mobile health application, and an extensive set of simulation experiments that assess the effectiveness of GOPRIME in terms of scalability, robustness and convergence speed.

Pulverization in Cyber-Physical Systems: Engineering the Self-Organizing Logic Separated from Deployment

“…Weyns and Georgeff [52] apply self-adaptation in a multi-agent system application for automated guided vehicles. An example of self-adaptive middleware architecture for dynamic service composition is described in [53], and GoPrime [54] offers a decentralized middleware for self-assembly of distributed services. A gossip protocol realizes decentralized data dissemination to maintain an assembly of services that fulfils global QoS goals.…”

Reinforcement Learning Techniques for Decentralized Self-adaptive Service Assembly

“…To cope with some of the challenges we have outlined in the introduction, this coordinated monitoring activity should scale with increasing system size, and be able to quickly react to changes occurring in the system (e.g., new offered services, variations of their quality). To this end, we adopt a gossip-based approach [1], which exploits epidemic protocols to achieve reliable information exchange among large sets of interconnected peers, also in presence of network volatility (e.g., peers join/leave the system suddenly). Specifically, in a gossip communication model, each peer in the system periodically exchanges information with a dynamically built peer set, and spreads information epidemically, similar to a virus in biological communities.…”

“…Instead, if S has a nonempty set of dependencies (S.Deps = ∅) and is not fully resolved, U t (S) is set to ⊥, i.e., the special value that is guaranteed to be "worse" than the quality of any fully resolved instance of S. Finally, if S has a nonempty set of dependencies and is fully resolved, U t (S) is computed using a function C : R (1+|S.Int|)m → R m , which combines the local load-dependent quality L(u(S), S.Out t ) with the overall quality of all S dependencies. The general equation (1) can be instantiated for specific quality attributes as described, for example, in [1]. Problem formalization -Our goal is to maximize the quality globally delivered by the services hosted in the system, ensuring at the same time fairness among services.…”

A Service-Oriented Middleware Enabling Decentralised Deployment in Mobile Multihop Networks

“…that is found in networks involving mobile devices, such as the IoT, the fog, etc. More flexible models and elastic tools were proposed by Researchers, like JavaCà&Là (JCL) [23], GoPrime [7], ParallelTheater [21], ActorEdge [2], and EmbJXTAChord [3], but in spite of their numerous good features, they often are Research tools designed to solve particular scientific problems, which make them hardly usable out-of-the-box in projects involving practical distributed computations. Adapting one of them would be a cumbersome work that has no guarantee of success as source codes, when they are fully available, are always very hard to embrace.…”