Abstract-Recent research in the different functional areas of modern routers have made proposals that can greatly increase the efficiency of these machines. Most of these proposals can be implemented quickly and often efficiently in software. We wish to use personal computers as forwarders in a network to utilize the advances made by researchers. We therefore examine the ability of a personal computer to act as a router. We analyze the performance of a single general purpose computer and show that I/O is the primary bottleneck. We then study the performance of distributed router composed of multiple general purpose computers. We study the performance of a star topology and through experimental results we show that although its performance is good, it lacks flexibility in its design. We compare it with a multistage architecture. We conclude with a proposal for an architecture that provides us with a forwarder that is both flexible and scalable.

Software routers based on Personal Computer (PC) architectures are receiving increasing attention in the research community. However, a router based on a single PC suffers from limited bus and Central Processing Unit (CPU) bandwidth, high memory access latency, limited scalability in terms of number of network interface cards, and lack of resilience mechanisms. Multi-stage architectures created by interconnecting several PCs are an interesting alternative since they allow to i) increase the performance of single-software routers, ii) scale router size, iii) distribute packet-forwarding and control functionalities, iv) recover from single-component failures, and v) incrementally upgrade router performance. However, a crucial issue is to hide the internal details of the interconnected architecture so that the architecture behaves externally as a single router, especially when considering the control and the management plane.In this paper, we describe a control protocol for a previously proposed multi-stage architecture based on PC interconnection. The protocol permits information exchange among internal PCs to support: i) configuration of the interconnected architecture, ii) packet forwarding, iii) routing table distribution, iv) management of the internal devices. The protocol is operating system independent, since it interacts with software routing suites such as Quagga and Xorp, and it is under test in our labs on a smallscale prototype of the multi-stage router.

Abstract-Recent research in PC based routers has proposed a distributed architecture. Such an architecture poses several challenges in the areas of scalability, robustness, efficiency of routing, latency and other issues. We examine the issue of decrease in throughput due to large routing tables in this architecture and propose partitioning as a solution. Our contribution is twofold: defining the concept of load for a node in a forwarding table trie and to show by simulation experiments the effectiveness of partitioning and its application to a distributed router.

WFQ is a widely used scheduler that enables QoS features in a router. The inherent centralized nature of the design of WFQ schedulers in most switches and routers creates several challenges when exported to distributed architectures. In this paper, we study the challenges of implementing WFQ in a distributed OSR, propose some novel techniques to address these challenges and compare the performance of our WFQ implementation in distributed OSR with that of a centralized WFQ scheme.