High-assurance service systems

Abstract

High-assurance systems (HAS) are information systems designed and implemented to achieve a degree of predictablebehavior, with predictability expressed in terms of their reliability, availability, safety, securityand timeliness(RASST) properties. High-assurance service systems(HASS) are a special class of HAS providing interactive, network-accessible and dynamically bound servicesto clients typically unknown at design time. Cyberphysical systems(CS) are, in turn, a special class of HASS responsible for automation and control services governing a wide range of physical processes. A service, in this context, results from transactional exchanges of information of specified valuebetween service providers (servers) and their customers (clients) on behalf of certain application-level objectives. These application-oriented transactions, carried out through discoverable service interface protocols, are governed by service level agreements(SLA) expressing performance-related assurancesthat servers agree, a priori,to provide to their clients. In dynamically bound service environments, specification of assurances depends on existence of a published set of performance indices and associated measurement processes for RASST and related properties. Consequently, high-assurance service systems require aperformance measurement framework(PMF) competent to express service-oriented value propositionsand their RASST dependencies. This chapter introduces a CS PMF, with a focus on three key elements. First, we introduce a cyberspatial reference model(CRM) for establishing the identity and location of distributed HASS servers and clients. Second, we define a set of service performance indices to measure RASST properties. Third, we develop an application neutral, yet operational definition of valueuseful in high assurance service systems for defining their respective value propositions. © 2009 Springer-Verlag US.