Evolving an adaptive industrial software system to use architecture-based self-adaptation

Abstract

Abstract-Although architecture-based self-adaptation has been widely used, there is still little understanding about the validity and tradeoffs of incorporating it into real-world software-intensive systems which already feature built-in adaptation mechanisms. In this paper, we report on our experience in integrating Rainbow, a platform for architecture-based self-adaptation, and an industrial middleware employed to monitor and manage highly populated networks of devices. Concretely, we reflect on aspects such as the effort required for framework customization and legacy code refactoring, performance improvement, and the impact of architecturebased self-adaptation on system evolution.

Decentralized self-adaptation in large-scale distributed systems

“…The great majority of studies on self-adaptive or autonomic systems have been evaluated through simulations or toy examples, without a complete empirical evaluation [18]. The main exception is the Rainbow framework [14], which has been used to convert a legacy system into a self-adaptive one in the work described in [9]. However, Rainbow addresses self-adaptation with a centralized approach, which may be not suitable for large scale distributed systems composed by a huge number of nodes.…”

Empirical resilience evaluation of an architecture-based self-adaptive software system

“…The Data Acquisition and Control Service (DCAS) [8] is a middleware from Critical Software that provides a reusable infrastructure to manage the monitoring of highly populated networks of devices. In particular, the middleware is designed to be seamlessly integrated with Critical's Energy Management System (csEMS)…”

“…The basic building blocks in a DCAS-based system ( Figure 1) are: 2 1 http://solutions.criticalsoftware.com/products services/csems/ 2 We herein consider a simplified version of the DCAS architecture. Further details about DCAS can be found in [8]. • Devices are equipped with one or more sensors to obtain data from the application domain (e.g., from wind towers, solar panels, etc.).…”

“…Contains 11 scenarios. Each scenario combines environmental stimulation (device delay) with one of the system changes identified 6 A detailed discussion of the functionality of the different sub-components of the processor node can be found in [8].…”

“…In particular, architecture-based self-adaptation [17,23,25] leverages architecture models to enable highlevel reasoning about the best way of adapting a system, and is regarded as a promising approach to building resilient software systems in a cost-effective manner. However, although architecture-based self-adaptation has been applied in practice and recent results indicate its potential benefits in terms of cost compared to embedded code-based adaptation mechanisms [8], there is still a lack of evidence supporting the arguments for its beneficial impact on system resilience [31].…”