Search

The advent of multicore CPUs has led to renewed interest in software routers built from commodity PC hardware. However fully exploiting the parallelism due to multiple cores requires the ability to efficiently parallelize the delivery of packets to cores. I.e. traffic arriving (departing) on an incoming (outgoing) link at a router is inherently serial and hence we need a mechanism that appropriately demultiplexes (multiplexes) traffic between a serial link and a set of cores. Recent efforts point to modern server network interface cards (NICs) as offering the required mux/demux capability through new hardware classification features. However there has been little evaluation of the extent to which these features match the requirements of software routing from the standpoint of both performance and functionality. This paper takes a first step towards such an evaluation comparing a commodity server NIC to both an idealized "parallel" NIC and a "serial only" NIC. We show that although commodity NICs do improve on serial only NICs they lag an ideal parallel NIC. We find similar gaps in the classification features these NICs offer. We thus conclude with recommendations for NIC modifications that we believe would improve their suitability for software routers.

International audience; Les processeurs multi-coeurs sur PC, en même temps que les progrès récents sur la performance de la mémoire et des bus, suggèrent un candidat idéal pour la construction d'architectures de routeurs purement logicielles qui soient flexibles et en même temps performantes. Dans cet article, nous présentons une architec- ture de commutation de paquets pour les routeurs logiciels ou virtuels dont l'objectif principal est de prévenir des accès mémoire coûteux tout en utilisant les ressources CPU disponibles au mieux pour traiter et commuter les paquets. Nous présentons premièrement le principe d'arbre de commutation dont la fonction est de gar- der les paquets commutés dans la même hiérarchie de cache du processeur. Nous identifions ensuite la source de contention principale des performances de commutation comme l'accès concurrent aux interfaces de sorties puis analysons les gains potentiels de performance en utilisant un verrou sur les interfaces de sortie ainsi qu'un mécanisme de retour sur l'état des queues du routeur. Finalement, nous proposons et évaluons une modification des arbres de commutation qui permet de contourner la contention due au partage des interfaces de sortie par les arbres.

Multi-core CPUs, along with recent advances in memory and buses, render commodity hardware a strong candidate for software router virtualization. In this context, we present the design of a new platform for virtual routers on x86 hardware. We also elaborate on our design choices in order to achieve both high performance and flexibility for packet processing., Electronic Communications of the EASST, Volume 17: Kommunikation in Verteilten Systemen 2009

We revisit the problem of scaling software routers motivated by recent advances in server technology that enable high speed parallel processing a feature router workloads appear ideally suited to exploit. We propose a software router architecture that parallelizes router functionality both across multiple servers and across multiple cores within a single server. By carefully exploiting parallelism at every opportunity we demonstrate a 35 Gbps parallel router prototype; this router capacity can be linearly scaled through the use of additional servers. Our prototype router is fully programmable using the familiar Click/Linux environment and is built entirely from off the shelf general purpose server hardware.

The advent of multicore CPUs has led to renewed interest in software routers built from commodity PC hardware. However, fully exploiting the parallelism due to multiple cores requires the ability to efficiently parallelize the delivery of packets to cores. I.e., traffic arriving (departing) on an incoming (outgoing) link at a router is inherently serial and hence we need a mechanism that appropriately demultiplexes (multiplexes) traffic between a serial link and a set of cores. Recent efforts point to modern server network interface cards (NICs) as offering the required mux/demux capability through new hardware classification features. However, there has been little evaluation of the extent to which these features match the requirements of software routing from the standpoint of both, performance and functionality. This paper takes a first step towards such an evaluation, comparing a commodity server NIC to both, an idealized "parallel" NIC and a "serial only" NIC. We show that although commodity NICs do improve on serial-only NICs they lag an ideal parallel NIC. We find similar gaps in the classification features these NICs offer. We thus conclude with recommendations for NIC modifications that we believe would improve their suitability for software routers.