%0 Journal Article %1 Milanovic2011 %A Milanovic, Nikola %A Milic, Bratislav %D 2011 %J IEEE TRANSACTIONS ON SERVICES COMPUTING %K %N 1 %T Automatic Generation of Service Availability Models %V 4

@article{Milanovic2011, added-at = {2014-08-13T13:47:29.000+0200}, author = {Milanovic, Nikola and Milic, Bratislav}, biburl = {https://www.bibsonomy.org/bibtex/268b83f030e23ef8ce6c4384ec1b97e96/junkerm}, file = {Milanovic2011.pdf:Milanovic2011.pdf:PDF}, groups = {public}, interhash = {1e55cf5c56c27ab3501d23e96376cf5d}, intrahash = {68b83f030e23ef8ce6c4384ec1b97e96}, journal = {IEEE TRANSACTIONS ON SERVICES COMPUTING}, keywords = {}, number = 1, quotations = {The existing methods and tools are powerful but difficult to apply directly for modeling service availability. Historical development of availability models has associated them with mission-critical systems that rarely change during their lifetime, such as real-time systems, avionics, or telecommunication circuits. The availability assessment procedures they offer are unable to adapt to fast-paced changes in modern SOA systems. Each time the IT infrastructure or business process changes, it is required to manually intervene and update, verify and evaluate availability model. This is the consequence of outdated philosophy behind standard availability models—they were primarily introduced to facilitate static fault-tolerant system design and follow the model !design!evaluate !forget (the model) process. Standard models are also very complex and manually built. Therefore, only isolated and relatively small portions of the enterprise infrastructure can be feasibly and realistically modeled. Furthermore, specialized and often highly mathematical knowledge and expertise is required to operate the tools, which is another reason for their low penetration in the service industry. Approaches addressing service availability almost exclusively investigate middleware concepts that can improve execution of existing services. As such, they are closer to the implementation of fault-tolerant systems (services), then to availability analysis, which is, strictly speaking, a necessary prerequisite for the fault-tolerant system design. In the majority of approaches, services are neither explicitly modeled nor their availability and the improvement thereof quantified. Service Availability Forum (www.saforum.org) is a consortium that develops high availability software interface specifications, specifically tailored to telecommunications sector. Many approaches introduce support for faulttolerant execution of single or composite Web services, using concepts such as transactional message delivery or message bus to ensure reliability, as in [9], [10], [11]. Other approaches use active and passive replication [12] or semantic substitution [13] to ensure fault-tolerance, but they all assess service availability only by simplified experiments. Other, nonstandard mechanisms for achieving fault-tolerance, such as use of generative communication and tuple space middleware, have been proposed [14]. Neither of these numerous approaches models reliability and availability explicitly or allows to quantify improvements gained using the proposed middleware concepts. The Web Service Reliable Messaging specification (WS-R) [15] is used to guarantee SOAP message delivery for Web services architecture, but does not provide a fault model or availability quantification. Some works expand WSReliableMessaging to include stochastic models for calculating optimal restart time for improving service performance, without generality or quantifiable availability assessment [16]}, timestamp = {2014-08-13T13:47:29.000+0200}, title = {Automatic Generation of Service Availability Models}, username = {junkerm}, volume = 4, year = 2011 }