Análise e implementação de técnicas de redução de tabelas de roteamento

Abstract

With the significant growth of communication systems, new strategies are being investigated in order to minimize memory resources for large-scale simulations. Specifically, this research considers routing tables, since their growth is one of the main scalability problems and has given rise to research seeking efficient solutions. Thus, this work identified, from a Systematic Literature Review (SLR), a set of routing table reduction techniques proposed in the literature and analyzed their feasibility of implementation for the context of IP routing. Thus, the research findings presented possibilities for advancing the integration between content-oriented/IP networks, addressing strategies initially aimed at the context of Named Data Networking (NDN), an architecture initially designed to meet the emerging needs of communication (Zhang; Burke, 2010). As a result, the Chatterjee et al. (2021) and Saxena e Raychoudhury (2016) strategies were selected, implemented and compared in terms of prefix reaggregation, memory usage and table reduction. From the analysis, it was possible to identify a reduction of the original table of 83.94% in the Chatterjee et al. (2021) strategy and 80.66% in the Saxena e Raychoudhury (2016) strategy. Additionally, a Mann-Whitney U statistical test was performed to verify the significant differences between the techniques in relation to execution time, memory usage and prefix reaggregation time. The results confirmed statistically significant differences, with the Chatterjee et al. (2021) strategy standing out for presenting shorter execution times and lower memory consumption, evidencing its superiority in terms of efficiency. Thus, the results indicate that the choice of technique directly impacts network performance, with the Chatterjee et al. (2021) strategy presenting significant reductions in storage. It is concluded that methods such as the Patricia Trie in conjunction with other methods, such as the Connected Dominating Set (CDS), offer potential solutions for minimizing routing table resources.