An analytical modeling framework to evaluate converged networks through business-oriented metrics Almir P. Guimarães a,b,n , Paulo R.M. Maciel b , Rivalino Matias Jr. c a Campus Arapiraca, Federal University of Alagoas, Arapiraca, AL, Brazil b Center of Informatics, Federal University of Pernambuco, Recife, PE, Brazil c School of Computer Science, Federal University of Uberlândia, Uberlândia, MG, Brazil article info Article history: Received 11 June 2012 Received in revised form 8 March 2013 Accepted 10 April 2013 Available online 27 April 2013 Keywords: Dependability Stochastic Petri Net Reliability Block Diagram Reliability importance Infrastructure design abstract Nowadays, society has increasingly relied on convergent networks as an essential means for individuals, businesses, and governments. Strategies, methods, models and techniques for preventing and handling hardware or software failures as well as avoiding performance degradation are, thus, fundamental for prevailing in business. Issues such as operational costs, revenues and the respective relationship to key performance and dependability metrics are central for defining the required system infrastructure. Our work aims to provide system performance and dependability models for supporting optimization of infrastructure design, aimed at business oriented metrics. In addition, a methodology is also adopted to support both the modeling and the evaluation process. The results showed that the proposed methodology can significantly reduce the complexity of infrastructure design as well as improve the relationship between business and infrastructure aspects. & 2013 Elsevier Ltd. All rights reserved. 1. Introduction The infrastructure required to support services of converged networks is increasingly more complex than ordinary commu- nication networks. It takes the form of complex system intercon- nections through a variety of network components such as routers, switches, firewalls, communication links, and other related components. In order to realize the design of converged networks, infra- structure metrics (e.g., availability, reliability, throughput) as well as their values should be selected, beyond the effort to achieve a design that minimizes costs and meets the selected metrics. Since converged networks are a means to helping the business meet its objectives, not formally taking business requirements into account is a major weakness. Ref. [10] explains that today's approach in meeting business requirements is to talk to business managers and ask them what the business needs in terms of availability, reliability, etc. These requirements are then expressed formally in a document called Service Level Agreement (SLA) [14], whereby the service provider makes certain guarantees that the service user can expect to be fulfilled. This problem should be considered seriously, since hundreds of thousands of dollars in financial outlay can easily be the difference between an ad hoc solution and another one that formally takes business considerations into account, even for medium-sized infrastructures [10]. In the last years, some optimization techniques have been proposed for system designs (telecommunication systems, manu- facturing systems and power systems) [5,7,9,11,15–17]. These systems consider reliability as an important design measure. Generally, the optimization techniques can be classified as linear programming, dynamic programming, integer programming, geometric programming, heuristic methods, genetic algorithms (GA) or a hybrid approach. Ref. [5] proposes a decomposition-based approach to exactly solve the multi-objective Redundancy Allocation Problem for series–parallel systems. The work in [7] describes and demon- strates two methods to intelligently reduce the size of the Pareto set. The first method is a pseudo-ranking scheme that helps the decision maker select solutions that reflect his/her objective function priorities. In the second approach, the authors used data mining clustering techniques to group the data by using the k- means algorithm to find clusters of similar solutions. Ref. [9] presents a method that can be used to design an ad hoc network topology optimized for higher reliability. Ref. [11] presents an algorithm based on a hybrid optimization approach (probabilistic solution discovery and Monte Carlo simulation) to solving pro- blems related to minimization of the network design cost, subject Contents lists available at SciVerse ScienceDirect journal homepage: www.elsevier.com/locate/ress Reliability Engineering and System Safety 0951-8320/$ - see front matter & 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ress.2013.04.008 n Corresponding author at: Campus Arapiraca, Federal University of Alagoas, Arapiraca, AL, Brazil. Tel.: +55 82 9331 5381. E-mail addresses: almir.guimaraes@arapiraca.ufal.br, almirpg@yahoo.com, almirguimaraes@yahoo.com.br (A.P. Guimarães), prmm@cin.ufpe.br (P.R.M. Maciel), rivalino@fc.ufu.br (R. Matias Jr.). Reliability Engineering and System Safety 118 (2013) 81–92