QoS-Aware Rule-Based Traffic-Efficient Multiobjective Service Selection in Big Data Space

The number of Web services has increased dramatically during the last few years. This has resulted in an increase in the volume of candidate services for tasks in composition systems. This has led to growth in the variety of nonfunctional properties in service selection, resulting in uncertainty (veracity issues) among such properties, which has severely affected the NP-hard aspects of service selection. Despite this, consumers in many areas would like access to a variety of selection methods such as linear programming and dynamic programming techniques. An additional problem is that the composition length (the number of tasks) of the workflow has increased, with the incorporation of research domains such as data science. These trending composition issues are challenging the computational power of existing methods. Such concerns have opened the door to research involving Big Data space. We propose a flexible, distributed selection algorithm that facilitates heterogeneous-selection methods to satisfy multiobjective composition requirements rather than rigid, specific composition requirements. However, service-selection processes in a Big Data space will inevitably increase traffic congestion caused by the increased volume of internal communication, particularly external traffic, such as Zipf and Pareto phenomena, and internal traffic during shuffling. To address these concerns, we propose solutions for each case. Our experiments demonstrate that the proposed traffic-efficient multiobjective method is well behaved when selecting services in Big Data space.