Possibilities for Improving Dependability of C4I2SR Systems Based on Service-Oriented Architecture

• Published in: 2009 Computation World: Future Computing, Service Computation, Cognitive, Adaptive, Content, Patterns pp. 315 - 324
• Main Authors: Jormakka, J., Lucenius, J.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.11.2009
• Subjects: Communication system control; Control systems; Force control; Internet; Logic testing; Military computing; Reconnaissance; Service oriented architecture; Strontium; Surveillance
• ISBN: 9781424451661; 1424451663
• DOI: 10.1109/ComputationWorld.2009.74

Service oriented architecture (SOA) is currently the way to design and integrate information services in enterprises. SOA is usually based on Web services,originating from the e-business world, but SOA is not only technology, it is more of a philosophy. SOA has been selected as the architectural solution for the new command, control, communication, computers, information, intelligence, surveillance and reconnaissance (C 4 I 2 SR) system of the Finnish defense forces. The task of the military is not Internet business and therefore SOA is not used for the purpose that it is created. Dependability demands, such as reliability, availability, security, of a C 4 I 2 SR system are high but a service oriented design approach cannot use many of traditional ways of achieving high dependability. In this paper we describe some dependability problems of SOA, and our ideas on how to solve these problems. Particularly, thorough design and management aspects are important. It seems like we have to model, design and test carefully both the underlying logic and protocols, and the high level services, i.e. combine a down-up and top-down approach. One aspect, which we have to be aware of are service interactions, at least if services get complicated. The non-functional requirements and performance parameters should be defined for a military SOA-based system. This may, however, be difficult in CC 4 I 2 SR as the final use cases are not necessarily predictable in advance. During use, the system should be monitored in order to identify the problems and enable correcting actions.