Multi-cloud service composition using Formal Concept Analysis

Abstract

Formal Concept Analysis is adapted to the multi-cloud service composition problem.Multi-cloud environment is modelled using lattice theory.Algorithms for optimal cloud combination are proposed.Experiments proved the ability of FCA to reduce the number of clouds and inter-cloud communication cost. Recent years have witnessed a rapid growth in exploiting Cloud environments to deliver various types of resources as services. To improve the efficiency of software development, service reuse and composition is viewed as a powerful means. However, effectively composing services from multiple clouds has not been solved yet. Indeed, existing solutions assume that the services participating to a composition come from a single cloud. This approach is unrealistic since the other existing clouds may host more suitable services. In order to deliver high quality service compositions, the user request must be checked against the services in the multi-cloud environment (MCE) or at least clouds in the availability zone of the user. In this paper, we propose a multi-cloud service composition (MCSC) approach based on Formal Concept Analysis (FCA). We use FCA to represent and combine information of multiple clouds. FCA is based on the concept lattice which is a powerful mean to classify clouds and services information. We first model the multi-cloud environment as a set of formal contexts. Then, we extract and combine candidate clouds from formal concepts. Finally, the optimal cloud combination is selected and the MCSC is transformed into a classical service composition problem. Conducted experiments proved the effectiveness and the ability of FCA based method to regroup and find cloud combinations with a minimal number of clouds and a low communication cost. Also, the comparison with two well-known combinatorial optimization approaches showed that the adopted top-down strategy allowed to rapidly select services hosted on the same and closest clouds, which directly reduced the inter-cloud communication cost, compared to existing approaches.