virtualization and sdn applications
play

Virtualization and SDN Applications 2 Virtualization Network - PDF document

4/1/2013 Virtualization and SDN Applications 2 Virtualization Network Virtualization Sharing physical hardware or software resources by Share physical network resources to form multiple multiple users and/or use cases diverse virtual


  1. 4/1/2013 Virtualization and SDN Applications 2 Virtualization Network Virtualization • Sharing physical hardware or software resources by • Share physical network resources to form multiple multiple users and/or use cases diverse virtual networks • Examples • Examples – Operating system shares physical hardware resources – Overlay and p2p networks – Virtual machine shares a physical machine with diverse – Virtual Private Networks (VPN) and multiple operating systems • Provide remote access to company’s network – Multiplexing shares a physical channel with multiple • Group remote computers in the same Virtual Local Area communication flows Network (VLAN). • Benefits: – Increases utilization of resources – Simplifies resource management 4 Network Virtualization Internal Network Virtualization • Properties of virtual switch • Two categories : – A virtual switch works – External network virtualization (most of this talk) much like a physical • Combining many networks, or parts of networks, into a Ethernet switch. virtual unit. – It detects which VMs are – Internal network virtualization logically connected to each • Providing network-like functionality to the software of its virtual ports and uses containers on a single system. that information to forward traffic to the correct virtual machines. 1

  2. 4/1/2013 Key properties of virtual network What are virtual networks used for? • Same purposes as non-virtualized networks without • Partitioning: each resource can be used interfering the operation of other virtual networks while concurrently by multiple VN instances sharing the key components among virtual networks • Isolation: the clear isolation of any VN from all – Coexistence of multiple VNs • Different VNs may use different network others technologies without interference • Abstraction: in which a given virtual resource need • Increase utilization not directly correspond to its component – Can support seamless migration/update of VNs resources – Can provide normalized set of interfaces and make it • Aggregation: aggregate multiple instances to easier to provision VNs obtain increased capabilities 7 8 Isolated “slices” Many operating systems, or The “Software-defined Network” Many versions 2. At least one good operating system 3. Well-defined open API Extensible, possibly open-source App App App App App App App App App App App Network Network Network Network Operating Operating Operating Operating Network Operating System System 1 System 2 System 3 System 4 Open interface to hardware 1. Open interface to hardware Virtualization or “Slicing” Layer Simple Packet Forwarding Open interface to hardware Hardware Simple Packet Forwarding Hardware Simple Packet Forwarding Hardware Simple Packet Simple Packet Forwarding Forwarding Hardware Hardware Simple Packet Simple Packet Forwarding Hardware Forwarding Simple Packet Hardware Forwarding Hardware Simple Packet Simple Packet Forwarding 9 Forwarding Hardware 10 Hardware Virtualized OpenFlow Substrate FlowVisor Creates Virtual Networks Larry’s Steve’s Dave’s Controller Larry’s Controller Steve’s Net Services Controller Net Services Net Services Dave’s Controller Controller Controller API API API OpenFlow Protocol OpenFlow OpenFlow Protocol Protocol ��������� FlowVisor ������ Network Hypervisor ��������� ������ & Policy Control FlowVisor slices OpenFlow OpenFlow networks, creating multiple ��������� ��������� Protocol ������ ������ isolated and programmable ��������� ��������� logical networks on the ������ ������ same physical topology. 12 11 2

  3. 4/1/2013 Slicing Policies Switch Based Virtualization Research VLAN 2 • The policy specifies resource limits for each Controller Flow Table slice: Research VLAN 1 Flow Table – Link bandwidth Controller Production VLANs Normal L2/L3 Processing – Maximum number of forwarding rules – Topology – Fraction of switch/router CPU – FlowSpace: which packets does the slice control? Use Case: VLAN Based Use Case: New CDN - Turbo Coral ++ Partitioning • Basic Idea: Build a CDN where you control the entire network • Basic Idea: Partition Flows based on Ports and – All traffic to or from Coral IP space controlled by Experimenter VLAN Tags – All other traffic controlled by default routing – Traffic entering system (e.g. from end hosts) is tagged – Topology is the entire network – VLAN tags consistent throughout substrate – End hosts are automatically added (no opt-in) Switch MAC MAC Eth VLAN IP IP IP TCP TCP Port src dst type ID Src Dst Prot sport dport Switch MAC MAC Eth VLAN IP IP IP TCP TCP Dave Port src dst type ID Src Dst Prot sport dport * * * * 1,2,3 * * * * * Turbo * * * * * 84.65.* * * * * Larry * * * * 4,5,6 * * * * * Coral * * * * * * 84.65.* * * * Steve * * * * 7,8,9 * * * * * Default * * * * * * * * * * FlowSpace: Maps Packets to Slices Use Case: Your Internet Protocol – A new layer 3 protocol – Replaces IP – Defined by a new Ether Type Switch MAC MAC Eth VLAN IP IP IP TCP TCP Port src dst type ID Src Dst Prot sport dport YourIP Your IP * * * * * * * * * Rest * * * !YourIP * * * * * * 3

  4. 4/1/2013 Dynamic Flow Aggregation on an OpenFlow Network Scope • Different Networks want different flow granularity (ISP, Backbone,…) • Switch resources are limited (flow entries, memory) • Network management is hard • Current Solutions : MPLS, IP aggregation How OpenFlow Helps? Applications of SDN • Dynamically define flow granularity by wildcarding arbitrary header fields • Granularity is on the switch flow entries, no packet rewrite or encapsulation • Create meaningful bundles and manage them using your own software (reroute, monitor) Higher Flexibility, Better Control, Easier Management, Experimentation 19 20 ElasticTree: Reducing Energy in Data Center Networks • Shuts off links and switches to reduce data center power • Choice of optimizers to balance power, fault tolerance, and BW • OpenFlow provides network routes and port statistics • The demo: • Hardware-based 16-node Fat Tree • Your choice of traffic pattern, bandwidth, optimization strategy • Graph shows live power openflow.org/videos and latency variation 21 22 demo credits: Brandon Heller, Srini Seetharaman, Yiannis Yiakoumis, David Underhill TutorialFlow Today’s Hands-On Session http://www.openflow.org/wk/index.php/OpenFlow_Tutorial Part 5 of OpenFlow Tutorial: http://www.openflow.org/wk/index.php/OpenFlow_Tutorial 23 24 4

  5. 4/1/2013 Tutorial Setup This talk wouldn’t be possible without: � Past slides from: Controller � Brandon Heller c0 port6633 � Nick McKeown � Rob Sherwood loopback OpenFlow Tutorial (127.0.0.1:6633) � Nick McKeown 3hosts-1switch � Rob Sherwood topology loopback s1 dpctl � Guru Parulkar (127.0.0.1:6634) OpenFlow Switch (user space � Srini Seetharaman process) � Yiannis Yiakoumis s1-eth0 s1-eth1 s1-eth2 � Guido Appenzeller � Masa Kobayashi h1-eth0 h3-eth0 h4-eth0 � Scott Shenker h2 h3 h1 � Sangjin Jeong 10.0.0.2 10.0.0.3 10.0.0.4 � others 25 virtual hosts 5

Recommend


More recommend