Adaptation of Web Services Based on QoS Satisfaction

Abstract

Requirements of Service Based Applications (SBAs) tend to change during the life cycle of the service. Therefore, adaptation and evolution of services become a necessity in order to provide the agreed Quality of Service (QoS) stated in a contract between the service provider and requestor. Recently, many adaptation methods have been proposed in the literature. However, there is no overall consensus in selecting the best strategy and the consequences of adaptation are usually neglected. In this paper, we propose an approach for service adaptation through defining a flexible service description using fuzzy parameters. This approach provides a compatibility mechanism that measures the aggregated satisfaction value of offered services to understand to what extent the quality changes are satisfiable according to the existing contract. According to the degree of satisfaction function we then propose our adaptation/evolution strategy.

Incremental service level agreements violation handling with time impact analysis

“…In [20], they proposed a cost-benefit model based on the value of change information (VOC) proposed in [19] to decide when the adaptation (recovery) should occur. In [40], a QoS satisfaction approach is proposed to support the decision on the appropriate adaptation (recovery) strategy. The work by [28] also proposed an adaptation (recovery) strategy selection approach.…”

“…Recent works e.g., [48], [40] has addressed a few issues in handling violation. As a result, several approaches have been proposed that include the aspect of handling patterns, the detection mechanism, the diagnosis mechanism, and the resolution mechanism.…”

Managing Expectations: Runtime Negotiation of Information Quality Requirements in Event-Based Systems

“…We refer to [19,25] for a detailed discussion due to space limitations and would like to focus on two related contributions: Kattepur et al [17] define a QoS metric similar to properties in expectations. However, they focus on interactions in heterogeneous SOA choreographies while expectations are information-centric; Pernici et al [21] use fuzzy parameters for deciding on web service adaptation. Those approaches, however, are based on direct contracts between service providers and service consumers, often assuming the existence of explicitly modeled workflows or a central authority for coordination.…”

“…As this comes at a cost, the middleware has to assess if the expectation should be satisfied or declined. Different optimization strategies can be applied to such a decision problem [21]. For example, we can apply a strategy aiming at pareto-optimal states for subscribers: we decline an adaptation to satisfy X e i ∈ Sat for subscriber i only if another expectation X e i is already satisfied for subscriber i (i.e., X e i ∈ Sat) and satisfying X e i would be more expensive than the current state; we decide to adapt in all other cases.…”

Comprehensive Explanation of SLA Violations at Runtime

“…For demonstration purposes, we have implemented a web 757 application 6 as a SALMonADA client in order to specify or 758 upload the WS-Agreement documents to monitor, execute 759 SALMonADA and receive the results. In this web appli-760 cation, we have introduced the WS-Agreements of ADA 761 and SALMon themselves.…”