RBSLA: Rule based service level agreements: RBSLA; knowledge representation for automated e-contract, SLA and policy management

The emergence of service oriented computing (SOC) together with novel business models such as On-Demand or Utility Computing leverage an open environment based upon loosely coupled and distributed services and dynamic service provider and service consumer relationships. Dynamic service supply chains (a.k.a. business services networks) based upon service-oriented and eventdriven architectures (SOAs and EDAs) and technology standards such as Web Services, Grid Computing and the Semantic Web ease the outsourcing of complex IT infrastructure to IT service providers. IT service providers must be able to ful l their policy and service-quality commitments based upon Service Level Agreements (SLAs) with their service customers. They need to manage, execute and maintain thousands of contracts for di erent customers and di erent types of services while facing rapidly changing business and system environments, a wide range of domain knowledge due to diverse organizational boundaries, huge amounts of scattered data managed in possibly distributed heterogeneous data sources and a great variety of more or less cooperative entities (human or automated agents) involved during the contracts life cycles. Moreover, they have to accommodate high requirements regarding correctness, robustness and traceability of drawn results and triggered reactions from the service contracts in order to establish trust and ful l legal compliance rules. While past research in service oriented computing has focused on the fairly static functional description and the operators of services including publication, discovery, selection, binding and composition of basic services the top layer in the SOA pyramid providing support for IT service management (ITSM) and in particular IT Service Level Management based upon SLAs describing non-functional service properties such as quality of service (QoS) and business policies has not been as thoroughly explored. I argue that the prevailing IT service level management tools and the actual SLA and contract languages such as WSLA, WS-Agreement or WS-Policy which mainly focus on the syntactical description of electronic contracts and policies using XML markup are not suitable to solve this task. The complexity of contractual logic in SLAs in an open, dynamic service oriented environment such as the Semantic Web requires new forms of knowledge representation and new technical solutions with a high degree of agility and exibility in order to e ciently engineer, manage and continuously monitor and enforce large amounts of complex and distributed IT service contracts. I propose a declarative rule-based approach for SLA representation and service level management (SLM) of IT services, where sophisticated logical knowledge representation (KR) concepts and arti cial intelligence (AI) techniques are used to describe the contracts in a generic way. Using rules for SLA representation has several advantages: 1. reasoning with rules is based on a semantics of formal logic a variation of rst order predicate logic enabling automated rule chaining by resolution and variable uni cation, which alleviates the burden of having to