Effective algorithms for consistent recovery in service-oriented systems

Funded by the National Science Centre: UMO-2011/03/D/ST6/01331 (2012-2016)

The service oriented systems, due to their loose-coupling and a great number of independent components that are susceptible to failures, are very error-prone. Failures are highly undesirable from the viewpoint of a service oriented system's clients, who expect that provided services and business processes are reliable. This makes the fault-tolerance an important aspect of the SOA applications. During the last years a lot of research has been carried out in the area of fault-tolerance. However, mechanisms known from the general distributed systems, among which are: replication, forward recovery in which the compensation actions are explicitly declared, and a rollback-recovery checkpoint-based approach, cannot be directly applied in SOA, due to specific characteristics of these systems. Thus, the existing fault-tolerance approaches have to be specially tailored, to address autonomy of service providers, a dynamic nature and longevity of the interactions, and the inherent constant interaction with the outside world, among the others.

Although some attempts to provide the fault-tolerance to the SOA systems have been undertaken, only the checkpointing and rollback-recovery mechanisms, which are the main focus of our work, allow in the case of failures of system entities, to automatically restore a failed system back to a consistent state, fully masking the failure occurrence at the same time. Unfortunately, this approach lacks proper theoretical foundations - the notion of a consistent state is ambiguous and has not been formalized in the context of SOA. Meanwhile, finding the consistent state of service-oriented computation is important for analyzing, testing or verifying properties of these computations. Thus, the lack of formally specified and recognized consistency requirements for SOA-compliant processing gravely prohibits the construction of provably correct rollback-recovery protocols. Therefore, the proposed project aimed at giving the necessary formal basis for any further in-depth research in this field. On the basis of proposed consistency definitions, the checkpointing and rollback-recovery protocols ensuring the appropriate consistency models were designed and fully tested in a real-life environment. Such protocols provide a convenient tool for the development of highly available and reliable SOA systems.

The results of our work will be also published in the form of scientific publications on conferences and in journal:

A. Kobusińska and D. Wawrzyniak, Maintaining replicated recovery log for RESTful services, International Journal of Grid and High Performance Computing (IJGHPC), vol. 7, no. 3, pp. 23?35, 2015

A. Kobusińska and D. Wawrzyniak, Replication of recovery log - an approach to enhance SOA reliability, in Distributed Applications and Interoperable Systems - 15th IFIP WG 6.1 International Conference, DAIS 2015, Held as Part of the 10th International Federated Conference on Distributed Computing Techniques (DisCoTec'15),Grenoble, France, June 2-4, 2015, Proceedings, pp. 152-157, 2015

A. Kobusińska, M. Holenko, and P. Zierhoffer, Handling consistent web service recovery in a RESTful way, in Proceedings of the 12th ACM International Conference on Computing Frontiers (CF'15), Ischia, Italy, May 18-21, 2015, pp. 39:1-39:2, 2015

A. Kobusińska and D. Wawrzyniak,Consistency of replicated log for service recovery,in Proceedings of 17th International Conference on Computational Science and Engineering (CSE'15),IEEE , Porto, Portugal, pp. 278-285, 2015

J. Brzeziński, M. Hołenko, A. Kobusińska, D. Wawrzyniak, and P. Zierhoffer,Towards relaxed rollback-recovery consistency in SOA, in Network and Parallel Computing (NPC'14), vol. 8707 of Lecture Notes in Computer Science, (Ilan, Taiwan), pp. 96-107, Berlin Heidelberg, Sept. 2014

J. Brzeziński and A. Kobusińska,On the equivalence of SOA rollback-recovery consistency models, in 25th IEEE International Symposium on Software Reliability Engineering Workshops (ISSRE'14) Workshops, Naples, Italy, November 3-6, 2014, RSDA'14,Naples, Italy), pp. 377-382, IEEE Computer Society, Nov. 2014

A. Danilecki, M. Holenko, A. Kobusinska, and P. Zierhoffer,The external recovery problem, in Euro-Par 2014: Parallel Processing Workshops (Euro-Par'14), Porto, Portugal, August 25-26, 2014, Revised Selected Papers, Part I, pp. 535-546, 2014

A. Danilecki, M. Hołenko, A. Kobusińska, and P. Zierhoffer,The impact of RESTful semantics on reducing the logging overhead in SOA, in Proc. of the 5th International Conference on Information, Intelligence, Systems and Applications (IISA'14), (Chania, Crete), pp. 228-233, IEEE Computer Society, June 2014

M. Hołenko, A. Kobusińska, D. Wawrzyniak, and P. Zierhoffer,The impact of service semantics on the consistent recovery in SOA, in Proc. of the 12th IEEE International Symposium on Parallel and Distributed Processing with Applications (ISPA'14), (Milano, Italy), pp. 109-116, IEEE Computer Society, Aug. 2014