| Summary: | The evolution of business software technologies is constant and is becoming increasingly complex which leads to a great probability of software/hardware failures. Business processes are built based on web services as they allow the creation of complex business functionalities. To attack the problem of failures presented by the use of web services, organizations are extrapolating the autonomic computing paradigm to their business processes as it enables them to detect, diagnose, and repair problems improving dependability. Sophisticated solutions that increase system dependability exist, however, those approaches have drawbacks; for example, they affect system performance, have high implementation costs, and or they may jeopardize the scalability of the system. To facilitate evolution to self-management, systems must implement the monitoring, analyzing, planning, and execution (MAPE) control loop. An open challenge for MAPE loop is to carry out in an efficient manner the diagnosis and decision-making processes, recollecting data from which the system can detect, diagnose, and repair potential problems. Also, dealt by systems dependability, specifically as fault tolerant mechanisms. One useful tool for this purpose is the communication induced checkpointing (CiC). We use CiC in attacking the dependability problem of using web services in a distributed and efficient manner. First, we present an approach for web services compositions that supports fault tolerance based on the CiC mechanism. Second, we present an algorithm aimed at web services compositions based on an autonomic computing and checkpointing mechanism. Experimental results support the feasibility of this concept proposal.
|
|---|
