Balanceamento de carga dinâmico, distribuído e centrado no conteúdo para Redes de Dados Nomeados

Abstract

In Named Data Networks (NDN), the name-based routing model predominantly benefits applications characterized by static content (i.e., content). Traditional NDN in-network caching is not suitable for dynamic content (i.e., services) as services are generated on demand. Furthermore, each service is also exclusive to each consumer, which lessens NDN in-network caching benefits. Conventional NDN routing strategies are also inadequate for an even service Interest distribution among the available replicas. In such scenarios, load balancing mechanisms can achieve better network utilization by maintaining workload balance among service providers' replicas. This work presents a load balancing strategy, called B3C, in which each router makes balancing decisions individually, considering their local information and without injecting control traffic. Simulation results show that B3C can evenly balance the load between the replicas and the traffic between the routes.