An instrumentation and control-based approach for distributed application management and adaptation

Distributed applications are notoriously difficult to develop and manage due to their inherent dynamics and heterogeneity of component technologies and network protocols. Middleware technologies dramatically simplify the development of distributed applications, but they still prove difficult to manage at runtime. This paper considers the "on-going" development of a framework that provides instrumentation and control services, which extend core middleware services, to realize the runtime management and adaptation of distributed applications. The instrumentation and control services are used in conjunction with dependency management utilities to measure performance, monitor behaviour and resolve the runtime inconsistencies and conflicts that may occur in distributed applications.