A multi-criteria network-aware service composition algorithm in wireless environments

Much work has been conducted to design effective and efficient algorithms for Quality of Service (QoS)-aware service algorithms in recent years. With the advent of high-capacity smart mobile devices and ubiquitous presence of wireless communications, service composition in wireless environments starts picking up momentum. While the majority of the know-how from service composition in wired fixed networks is still applicable to service composition in wireless environments, wireless networks also impose their own features that need to be specifically taken into consideration during service engineering process. To this end, this paper proposes a network-aware algorithm for service composition where network features such as network availability and delay are considered during service composing process. These network metrics incorporate with other non-network QoS parameters such as price and reputation, that are usually utilized in service composition in the current literature. Together these metrics specify the QoS constraints that need to be satisfied by a successful service composition, which is formulated as a multi-dimension multi-choice 0–1 knapsack problem (MMKP) in this paper. The evaluation results show the effectiveness of the proposed algorithm and also indicate the impact of the network parameters such as delay on the performance of service composition.