Cluster ComputingEngineeringHigh Performance Computer NetworkComputer ArchitectureNetwork AnalysisInterconnection Network ArchitectureData Center NetworkTree HierarchySystems EngineeringParallel ComputingNetwork DesignData Center SystemComputer EngineeringInterconnection NetworkData CentersData Center NetworkingData Center NetworksData Center ManagementServer MachinesEdge ComputingCloud Computing
Data‑center networking seeks to interconnect many servers cost‑effectively, with high balanced capacity and fault tolerance, yet the prevailing tree‑switch hierarchy cannot satisfy these requirements. This work proposes a new server‑interconnection architecture that actively uses both built‑in Ethernet ports on commodity servers to eliminate the need for expensive high‑level switches. The FiConn design employs both server ports and low‑end commodity switches to form a scalable topology, with a routing scheme that balances link levels and adapts to traffic state for efficient utilization. FiConn is proven scalable to hundreds of thousands of servers with low diameter and high bisection width, and simulations show its routing mechanisms achieve high throughput.
The goal of data center networking is to interconnect a large number of server machines with low equipment cost, high and balanced network capacity, and robustness to link/server faults. It is well understood that, the current practice where servers are connected by a tree hierarchy of network switches cannot meet these requirements (Fares et al., 2008 and Guo et al., 2008). In this paper, we explore a new server-interconnection structure. We observe that the commodity server machines used in today's data centers usually come with two built-in Ethernet ports, one for network connection and the other left for backup purpose. We believe that, if both ports are actively used in network connections, we can build a low-cost interconnection structure without the expensive higher-level large switches. Our new network design, called FiConn, utilizes both ports and only the low-end commodity switches to form a scalable and highly effective structure. Although the server node degree is only two in this structure, we have proven that FiConn is highly scalable to encompass hundreds of thousands of servers with low diameter and high bisection width. The routing mechanism in FiConn balances different levels of links. We have further developed a low-overhead traffic-aware routing mechanism to improve effective link utilization based on dynamic traffic state. Simulation results have demonstrated that the routing mechanisms indeed achieve high networking throughput.
19
Sanjay Ghemawat · Communications of the ACM · 2008 · 18.4K citations · Full text
Sanjay Ghemawat, Howard Gobioff, Shun-Tak A. Leung · ACM SIGOPS Operating Systems Review · 2003 · 5K citations
William J. Dally, Brian Towles · 2001 · 3.2K citations
Engineering, Network Routing, On-chip Interconnection Networks +20
Michael Isard, Mihai Budiu, Yuan Yu et al. · 2007 · 2.4K citations