11 January 2025 in Computer Communications
20 July 2023 in Journal of Network and Systems Management
Autonomous network management is crucial for Fifth Generation (5G) and Beyond 5G (B5G) networks, where a constantly changing environment is expected and network configuration must adapt accordingly. Modeling tools are required to predict the impact on performance (packet and delay loss) when new traffic demands arrives and when changes in routing paths are applied in the network. Mathematical analysis and network simulators are techniques for modeling networks but both have limitations, as the former provides low accuracy and the latter requires high execution times. To overcome these limitations, machine learning (ML) algorithms, and more specifically, graph neural networks (GNNs), are proposed for network modeling due to their ability to capture complex relationships from graph-like data while predicting network properties with high accuracy and low computational requirements. However, one of the main issues when using GNNs is their lack of generalization capability to larger networks, i.e., when trained in small networks (in number of nodes, paths length, links capacity), the accuracy of predictions on larger networks is poor. This paper addresses the GNN generalization problem by the use of fundamental networking concepts. Our solution is built from a baseline GNN model called RouteNet (developed by Barcelona Neural Networking Center-Universitat Politècnica de Catalunya (BNN-UPC)) that predicts the average delay in network paths, and through a number of simple additions significantly improves the prediction accuracy in larger networks. The improvement ratio compared to the baseline model is 101, from a 187.28% to a 1.828%, measured by the Mean Average Percentage Error (MAPE). In addition, we propose a closed-loop control context where the resulting GNN model could be potentially used in different use cases.
21 October 2014 in IEEE Communications Letters
As a potential solution to the compact routing problem, geometric routing has been proven to be both simple and heuristically effective. These routing schemes assign some (virtual) coordinates in a metric space to each network vertex through the process called embedding. By forwarding packets to the closest neighbor to the destination, they ensure a completely local process with the routing table bounded in size by the maximum vertex degree. In this letter, we present an embedding of any finite connected graph into a metric space generated by algebraic groups, and we prove that it is greedy (guaranteed packet delivery). Then, we present a specialized compact routing scheme relying on this embedding for scale-free graphs. Evaluation through simulation on several Internet topologies shows that the resulting stretch remains below the theoretical upper bounds.
23 June 2017 in IEEE Network
A wide range of social, technological and communication systems can be described as complex networks. Scale-free networks are one of the well known classes of complex networks in which nodes' degrees follow a power-law distribution. The design of scalable, adaptive and resilient routing schemes in such networks is very challenging. In this article we present an overview of required routing functionality, categorize the potential design dimensions of routing protocols among existing routing schemes, and analyze experimental results and analytical studies performed so far to identify the main trends/trade-offs and draw main conclusions. Besides traditional schemes such as hierarchical/shortest-path path-vector routing, the article pays attention to advances in compact routing and geometric routing since they are known to significantly improve scalability in terms of memory space. The identified trade-offs and the outcomes of this overview enable more careful conclusions regarding the (un-)suitability of different routing schemes to large-scale complex networks and provide a guideline for future routing research.