Predictable Efficiency for Reconfiguration of Service-Oriented Systems with Concerto

Abstract

Dynamic reconfiguration of distributed software systems is nowadays gaining interest because of the emergence of dynamic IoT and smart applications as well as large scale dynamic infrastructures (e.g., Fog and Edge computing). When quality of service and experience is of prime importance, efficient reconfiguration is necessary, as well as performance predictability to decide when a reconfiguration should occur. This paper tackles the problem of efficient execution of a reconfiguration plan and its predictability with Concerto, a reconfiguration model supporting a high level of parallelism. Evaluation performed on synthetic cases and on two real production scenarios show that Concerto provides better performance than state-ofthe-art systems with accurate time estimation.

Component-based Distributed Software Reconfiguration:A Verification-oriented Survey

“…Components are peers that can interact with few restrictions. Examples of such models include CCM [108], CCA [23], L 2 C [25,48], BIP [20,26], Concerto [41], Aeolus [54], Coqcots [32], and actor models [67].…”

“…Consequently, many models also distinguish between active and inactive components [25,31]. Finally, other models, specific to the management aspect of component-based architectures (comparable to DevOps tools) offer either advanced fixed life-cycle definitions [59,60], or the definition of a programmable life-cycle with an arbitrary number of states [41,42,54]. Transitions between these states often have to be specified and mapped to actions, for example with a finite state machine [54,59].…”

SMT-Based Planning Synthesis for Distributed System Reconfigurations

“…These three commands allow parallel asynchronous execution in Concerto, leading to more efficient reconfigurations. Based on the description of the components provided by their developers, Concerto can execute reconfiguration scripts , allowing for empirical performance comparisons [10].…”

“…However, if multiple solutions are returned, the best possible sequence is picked, according to some (possibly user-defined) selection criterion. Some interesting optimization criteria are: the length of a sequence, its execution time (if time estimations are available for individual transitions, this may be computed with great accuracy [10]), the number of transitions it executes sequentially, or the number of ports it (de)activates. In our experiments, we used this last criterion, as it picks the component reconfiguration that is least likely to induce changes in other components, leading to simpler and potentially faster reconfiguration plans.…”

Toward safe and efficient reconfiguration with Concerto

“…Concerto has previously been presented in [21], focusing on its performance estimation capabilities, but with only an overview of the model itself. In this article, we now provide a thorough description of the semantics of Concerto and extend the experiments to measure the performance benefits of its execution model.…”