Jean Tourrilhes

DevoFlow: scaling flow management for high-performance networks

OpenFlow is a great concept, but its original design imposes excessive overheads. It can simplify network and traffic management in enterprise and data center environments, because it enables flow-level control over Ethernet switching and provides global visibility of the flows in the network. However, such fine-grained control and visibility comes with costs: the switch-implementation costs of involving the switchs control-plane too often and the distributed-system costs of involving the OpenFlow controller too frequently, both on flow setups and especially for statistics-gathering. In this paper, we analyze these overheads, and show that OpenFlows current design cannot meet the needs of high-performance networks. We design and evaluate DevoFlow, a modification of the OpenFlow model which gently breaks the coupling between control and global visibility, in a way that maintains a useful amount of visibility without imposing unnecessary costs. We evaluate DevoFlow through simulations, and find that it can load-balance data center traffic as well as fine-grained solutions, without as much overhead: DevoFlow uses 10--53 times fewer flow table entries at an average switch, and uses 10--42 times fewer control messages.OpenFlow is a great concept, but its original design imposes excessive overheads. It can simplify network and traffic management in enterprise and data center environments, because it enables flow-level control over Ethernet switching and provides global visibility of the flows in the network. However, such fine-grained control and visibility comes with costs: the switch-implementation costs of involving the switchs control-plane too often and the distributed-system costs of involving the OpenFlow controller too frequently, both on flow setups and especially for statistics-gathering. In this paper, we analyze these overheads, and show that OpenFlows current design cannot meet the needs of high-performance networks. We design and evaluate DevoFlow, a modification of the OpenFlow model which gently breaks the coupling between control and global visibility, in a way that maintains a useful amount of visibility without imposing unnecessary costs. We evaluate DevoFlow through simulations, and find that it can load-balance data center traffic as well as fine-grained solutions, without as much overhead: DevoFlow uses 10--53 times fewer flow table entries at an average switch, and uses 10--42 times fewer control messages.

DevoFlow: cost-effective flow management for high performance enterprise networks

The OpenFlow framework enables flow-level control over Ethernet switching, as well as centralized visibility of the flows in the network. OpenFlows coupling of these features comes with costs, however: the distributed-system costs of involving the OpenFlow controller on flow setups, and the switch-implementation costs of involving the switchs control plane too often. In this paper, we analyze the overheads, and we propose DevoFlow, a modification of the OpenFlow model in which we try to gently break the coupling between centralized control and centralized visibility, in a way that maintains a useful amount of visibility without imposing unnecessary costs.

CORONET: Fault tolerance for Software Defined Networks

Software Defined Networking, or SDN, based networks are being deployed not only in testbed networks, but also in production networks. Although fault-tolerance is one of the most desirable properties in production networks, there are not much study in providing fault-tolerance to SDN-based networks. The goal of this work is to develop a fault tolerant SDN architecture that can rapidly recover from faults and scale to large network sizes. This paper presents CORONET, a SDN fault-tolerant system that recovers from multiple link failures in the data plane. We describe a prototype implementation based on NOX that demonstrates fault recovery for emulated topologies using Mininet. We also discuss possible extensions to handle control plane and controller faults.

Robust broadcast: improving the reliability of broadcast transmissions on CSMA/CA

This paper presents a scheme to improve the efficiency of radio MAC protocols in the case of broadcast and multicast transmissions, like TCP/IP multicasting. First, the reliability problems with broadcast packets and their consequences are analysed. Then the robust broadcast scheme is presented, which decreases the probability of loss of broadcast packets over MAC protocols based on CSMA/CA. Finally, the new protocol is simulated against other simple solutions to show how it performs.

meSDN: mobile extension of SDN

Mobile devices interact wirelessly with a growing proliferation of cloud-based applications. Due to significant traffic growth and a wide variety of multimedia solutions, enterprise IT departments are demanding more fine-grained visibility and control of mobile traffic. They want to deliver optimal performance and a high quality of experience to a variety of users and applications. In the wired world, Software-Defined Networking (SDN) is a technology being embraced to deliver performance guarantees to end users by dynamically orchestrating quality of service (QoS) policies on edge switches and routers. Guaranteeing performance in a wired access network does not require any network control on clients, because the last hop between the network edge and wired device is a dedicated point-to-point link (e.g. Ethernet). However, this is not the case with wireless LANs (WLAN), since the last hop is a shared half-duplex medium and the WiFi MAC protocol does not allow access points to coordinate client uplink transmissions or 802.11 QoS settings. Hence, we argue that the SDN paradigm needs to be extended to mobile clients to provide optimal network performance between the cloud and wirelessly-connected clients. In this paper, we propose a framework called meSDN and demonstrate that it enables WLAN virtualization, application-aware QoS and improves power-efficiency from our prototype on Android phones.

Fragment adaptive reduction: coping with various interferers in radio unlicensed bands

This paper studies the impact of common interferers found in unlicensed bands on 802.11 and other CSMA systems. First, we describe those interferers and the MAC techniques used to combat them. Then, we present fragment adaptive reduction, a scheme to overcome the effect of non-CSMA repetitive interferers, and its implementation in the SWAP protocol. We also describe a simple model of domestic microwave oven interference. We finish by some exhaustive simulations of the impact of microwave oven and other interferers on 802.11, which illustrate the benefits of the fragment reduction scheme.

Contact networking: a localized mobility system

MobileIP, the standard for Internet mobility, enables transparent mobility for a mobile node, but requires communication to take a multihop path through the nodes Home Agent. Although a user with a multiple-interface mobile node may desire the ability to communicate locally, perhaps while disconnected from the Internet, MobileIP offers no such support.Contact Networking provides lightweight, localized network communication to a node with diverse network interfaces. The goal is to provide support for local connectivity equivalent to that provided by MobileIP for remote connectivity. The concept of link-layer awareness enables Contact Networking to tailor its operation to different links, using link-layer native services to implement abstract services when possible. Interface management and autoconfiguration insulate the user from concerns about the number and type of interfaces available.In this paper, we motivate the need for localized mobility, and present the design and architecture of Contact Networking. Details of our prototype implementation illustrate the complexities of providing a localized mobility facility.

Ensemble routing for datacenter networks

This paper describes Hash-Based Routing (HBR), an architecture that enhances Ethernet to support dynamic management for multipath networks in scalable datacenters. This work enhances HBR to support flow ensemble management for large-scale networks of arbitrary topology. Ensemble routing eliminates measurement and control for individual flows and instead manages using summary data thus providing a unique capability for reactive datacenter-wide network management. HBR provides seamless interoperability with Ethernet and supports the attachment of unmodified L2 hosts and devices including FCoE devices within converged fabrics. Simulation experiments demonstrate efficient multipath routing for a variety of scalable topologies. Optimized routing maintains efficiency in the presence of network faults and implements spatial Quality of Service to dynamically provision physical network hardware among co-hosted tenants or applications.

Killer Fabrics for Scalable Datacenters

Most datacenter networks are based on specialized edge-core topologies, which are costly to build, difficult to maintain and consume too much power. We propose enhancements to layer-two (L2) Ethernet switches to enable multipath L2 routing in scalable datacenters. This replaces an expensive router with commodity switches. Our hash-based routing approach reuses and minimally extends hardware structures in high-volume switches, while exposing a powerful network management inter-face for multipath load balancing, QoS differentiation, and resilience to faults. Simulation results demonstrate near-optimal load balancing for uniform and non-uniform traffic patterns, and effective management of large datacenter networks independent of the number of traffic flows.

Co-link configuration: using wireless diversity for more than just connectivity

This paper describes a novel approach to setting up peer to wireless links, such as IrDA, Bluetooth, and 802.11b. We explain how the lack of physical contact inherent in wireless creates new configuration challenges and ease of use issues, and we discuss the limitations of current wireless autoconfiguration techniques. We present co-link configuration and co-link activation, two techniques using wireless diversity to improve the performance of minimise the power consumption of autoconfiguration. We then describe the implementations of co-link, the simple IrDA connection point and the peer-to-peer co-link implemented within the connection diversity framework.