No abstract

We investigate dynamic decision mechanisms for composite web services maximizing the expected revenue for the providers of composite services. A composite web service is represented by a (sequential) workflow, and for each task within this workflow, a number of service alternatives may be available. These alternatives offer the same functionality at different price-quality levels. After executing a sub-service, it is decided which alternative of the next sub-service in the workflow is invoked. The decisions optimizing expected revenue are based on observed response times, costs and responsetime characteristics of the alternatives as well as end-toend response-time objectives and corresponding rewards and penalties. We propose an approach, based on dynamic programming, to determine the optimal, dynamic selection policy.Extensive numerical examples show significant potential gain in expected revenues using the dynamic approach compared to other, non-dynamic approaches.

Abstract-In this paper we investigate sequential decision mechanisms for composite web services. After executing each sub-service within a sequential workflow, decisions are made whether to terminate or continue the execution of the workflow. These decisions are based on observed response times, expected rewards, and typical Service Level Agreement parameters such as costs, penalties, and agreed response-time objectives. We propose a model for the sequential decision-making process within which we explore a couple of decision algorithms. We benchmarked these algorithms against the profit made when executing the workflow without decision-making. We show that algorithm based on backward recursion principle of dynamic programming is optimal with respect to profit. Next, we analyse the structure of erroneous decisions for both algorithms and show that significant profit gains can be obtained by sequential decision making.

No abstract

Next-generation service offerings will be increasingly based upon combining and integrating information from multiple logically and geographically distributed servers, interconnected by communication networks. Different administrative domains own these servers and networks. For the commercial success of these services, it is important for service providers (SPs) to predict and control the end-to-end Quality-of-Service (QoS) perceived by the end users. We focus on transactionbased services, such as E-business applications, for which control of end-to-end response and download times determine customer satisfaction. Today, no mature solutions exist for the problem of realizing high and guaranteed end-to-end QoS for transaction-based services in multi-domain environments. Service Level Agreements (SLAs) are a well-recognized concept to obtain QoS guarantees at the network level. However, in the context of transaction-based services both server and network domains need to be taken into account. Furthermore, currently no satisfactory solutions exist for SPs to determine the set of combinations of per-domain SLAs that they need to negotiate with the other domain owners to deliver the desired end-to-end QoS. To this end, in this paper we introduce the new concept called SLA negotiation space, i.e. the set of combinations of perdomain SLAs that SPs need to negotiate with other domain owners to realize desired end-to-end QoS levels. In addition, to identify the SLA negotiation space, we propose a modelling framework and a step-by-step approach to quantify the complex relation between the per-domain SLA parameters and the end-to-end QoS. A specific feature of our modelling framework is that it explicitly incorporates the SLA parameters, which has not been proposed before. The practical usefulness of our results is demonstrated by a realistic example.