Author | Conference | Journal | Organization | Year | DOI
Look for results that meet for the following criteria:
Academic
Publications
VL2: a scalable and flexible data center network

VL2: a scalable and flexible data center network,10.1145/1592568.1592576,Albert G. Greenberg,James R. Hamilton,Navendu Jain,Srikanth Kandula,Changhoon

Edit
VL2: a scalable and flexible data center network   (Citations: 115)
BibTex | RIS | RefWorks Download
To be agile and cost effective, data centers should allow dynamic resource allocation across large server pools. In particular, the data center network should enable any server to be assigned to any service. To meet these goals, we present VL2, a practical network architecture that scales to support huge data centers with uniform high capacity between servers, performance isolation between services, and Ethernet layer-2 semantics. VL2 uses (1) flat addressing to allow service instances to be placed anywhere in the network, (2) Valiant Load Balancing to spread traffic uniformly across network paths, and (3) end-system based address resolution to scale to large server pools, without introducing complexity to the network control plane. VL2's design is driven by detailed measurements of traffic and fault data from a large operational cloud service provider. VL2's implementation leverages proven network technologies, already available at low cost in high-speed hardware implementations, to build a scalable and reliable network architecture. As a result, VL2 networks can be deployed today, and we have built a working prototype. We evaluate the merits of the VL2 design using measurement, analysis, and experiments. Our VL2 prototype shuffles 2.7 TB of data among 75 servers in 395 seconds – sustaining a rate that is 94% of the maximum possible.
Conference: ACM SIGCOMM Conference - SIGCOMM , pp. 51-62, 2009
Cumulative Annual
View Publication
The following links allow you to view full publications. These links are maintained by other sources not affiliated with Microsoft Academic Search.
More »
    • ...At core routers oversubscription is normally much higher, 1:20 is not unusual [15] and can be as high as 1:240 [26].,These issues have led to proposals for new network topologies for data center clusters [11, 26, 27, 40], which aim at increasing the bandwidth available by removing network oversubscription.,There have also been several proposals to improve the network topologies in data centers, including switchbased topologies [11, 26, 28, 40] and direct-connect (or hybrid) topologies [9, 27, 39]...

    Paolo Costa, et al. Camdoop: Exploiting In-network Aggregation for Big Data Applications

    • ...Previous work on designing network topologies [16, 29] and resource allocation mechanisms [6, 17] focused on tolerating failures of network components by providing multiple paths between server pairs and on ensuring predictable network behavior through better bisection bandwidth and reservation...

    Peter Bodík, et al. Surviving failures in bandwidth-constrained datacenters

    • ...Full-mesh COREs refer to the full-mesh interconnections between COREs and containers, i.e., every container connects to every core switch [10,21].,First, there are far more flows than the diversity of paths in DCNs [10, 17].,Second, hash-based flow-level load balancing is widely used at the link level, switch level, and path level in production DCNs [10,12].,Our intuition is that hashing many flows onto a relatively small number of paths leads to even load balancing [10]...

    Xin Wu, et al. NetPilot: automating datacenter network failure mitigation

    • ...However, recently developed CLOS networks [16, 15, 24]—large numbers of small commodity switches with redundant interconnections—have made it economical to build non-oversubscribed full bisection bandwidth networks at the scale of a datacenter for the first time.,Our FDS testbed uses a two-layer CLOS network [15, 16], which in its largest configuration consists of 8 “spine” routers and 14 “TORs” (Top-Of-Rack routers).,PortLand [24] and VL2 [15] make it economically feasible to build datacenter-scale full bisection bandwidth networks...

    Edmund B. Nightingale, et al. Flat Datacenter Storage

    • ...Even in a network with no oversubscription [43,44], Bazaar-I is able to accept 10% more requests and improves the goodput by 27% relative to Baseline...

    Virajith Jalaparti, et al. Bazaar: Enabling Predictable Performance in Datacenters

Sort by:

Citations (115)

Share this on