An Effort Estimation Method for Service-Oriented Architecture

Abstract

Determining size and effort of SOA systems is critical for managing SOA projects. As a consequence, a number of methods have been proposed to estimate effort of building SOA projects but the problem of estimating SOA development effort still remains largely unresolved mainly because there is limited attempt in using size metrics to estimate SOA development effort. To address this problem, this study proposed an effort estimation method for SOA centred on size metrics and effort factors. The proposed method enables estimation of effort factors using fuzzy logic technique to improve on estimation accuracy. The method was automated into a tool to facilitate entry of parameters and display of results. The study employed experiment research design based on 15 SOA projects developed by computer science undergraduate students to validate the proposed estimation method. To complement the experiment, we used a survey study involving 20 programmers from the industry to confirm the relevance of effort estimation factors proposed in this study. Result from the experiment revealed that the proposed method is more accurate and returned a lower Mean magnitude of relative error (MMRE). Response from the survey showed that the proposed effort factors are valid and they have influence on SOA development effort.