Assessing Computer Network Efficiency Using Data Envelopment Analysis and Multicriteria Decision Analysis Techniques

Measuring computer network efficiency is of great importance for a network manager. When a detailed picture of the current state of the network is available, we can use analytical and simulated-based methods. Instead, it is more difficult to measure network efficiency, when the details of network topologies and internal characteristics are not available. In this paper, we present operations research techniques, which have been used in measuring the efficiency of economic systems, namely data envelopment analysis (DEA), stochastic frontier analysis (SFA) and preference ranking organization method for enrichment evaluations (PROMETHEE). We investigate whether these techniques can be used to assess network efficiency and provide the network manager with a complete set of quantitative suggestions towards relative efficiency improvement of the networks determined as inefficient. The average efficiencies extracted in our benchmark from DEA and SFA are almost equal. Furthermore, the decision maker can adjust the weights of the criteria in PROMETHEE in order to build different scenarios. Consequently, we claim that these methods can be used for measuring computer network efficiency, when networks' internal characteristics are not available. Copyright (C) 2014 John Wiley & Sons, Ltd.