Title: Distributed multi-user QoS-aware service selection
Authors: Adrian Satja Kurdija; Marin Šilić; Goran Delač; Klemo Vladimir
Addresses: Faculty of Electrical Engineering and Computing, University of Zagreb, Zagreb, Croatia ' Faculty of Electrical Engineering and Computing, University of Zagreb, Zagreb, Croatia ' Faculty of Electrical Engineering and Computing, University of Zagreb, Zagreb, Croatia ' Faculty of Electrical Engineering and Computing, University of Zagreb, Zagreb, Croatia
Abstract: Previous research on service selection in cloud environments often assumes a centralised broker that keeps track of all candidate services and selects the appropriate candidate for each user request based on quality of service (QoS). As the number of users and services increases, the centralised service selection algorithm becomes a bottleneck. We propose a distributed, multi-broker selection algorithm that enables real-time processing of a large number of user requests. Each broker keeps track of a subset of users and a subset of service candidates, and rebalancing is performed to ensure that no broker is overloaded. The selection procedure considers the user QoS requirements, service QoS values and their processing capacities. Simulations demonstrate a clear advantage of the proposed approach over a single-broker approach by up to 96% in terms of QoS satisfaction when using 100 brokers, and especially in terms of response time satisfaction, showing positive impact of user and service balancing, and analysing various selection algorithms run by the brokers.
Keywords: service selection; quality of service; QoS; distributed systems; multi-broker; load balancing; services computing.
DOI: 10.1504/IJWGS.2022.126120
International Journal of Web and Grid Services, 2022 Vol.18 No.4, pp.411 - 436
Received: 29 May 2021
Accepted: 03 Apr 2022
Published online: 11 Oct 2022 *