Self-Management of COTS Component-Based Systems Using Wrappers

Abstract

This paper describes an approach to designing COTS (Commercial off-the-Shelf) components-based software system in which each COTS component is self-managed by a wrapper to improve dependability of systems. The wrapper for self-managing a COTS component detects the anomalies of the component; it reconfigures the anomalous component at run-time; and it repairs the anomalous component. The wrappers co-operate eachother to reconfigure COTS components, so that they make the system more reliable. The approach suggested in this paper is applied to the distributed elevator system with multiple elevators.

Dependability for high-tech systems: an industry-as-laboratory approach

“…Related to our approach is a method to wrap COTS components and monitor them using specifications expressed as a UML state diagrams [16]. Other related work consists of assertion-based approaches such as run-time verification [4].…”

New Design Techniques for Enhancing Fault Tolerant Cots Software Wrappers

“…In general it is necessary to use 'middleware', possibly based on standard infrastructure technologies to integrate the inevitably disparate components (May 2002). This middleware can also be used to play the role of a safety wrapper, and offers an important opportunity to include component adaptation and monitoring strategies, to help ensure the overall system's dependability (Shin and Paniagua 2006). One significant use of these wrappers is to enable the replacement one COTS software component of a safety critical system by another without significantly modifying the wrapper itself: for example, an upgraded component.…”

Model-Based Run-Time Error Detection

“…Related work on adding a control loop to an existing system is described in the middleware approach of [12] where components are coupled via a publish-subscribe mechanism. A method to wrap COTS components and monitor them using specifications expressed as a UML state diagrams is presented in [13]. The analogy between self-controlling software and control theory has already been observed in [14].…”