Download presentation
Presentation is loading. Please wait.
1
Network virtualization
Zilong Ye, Ph.D.
2
What is Virtualization?
2 Transparent abstraction of computing platform and resources Multiple logical interpretations of the physical characteristics Virtualized everything Virtual machines: VMware, Xen Storage virtualization: SAN Data-center virtualization What is Virtualization?
3
Network Virtualization for Dummies
3 Making a physical network appear as multiple logical ones Physical Network Virtualized Network - 1 Virtualized Network - 2
4
Why network virtualization
Why Virtualize ? 4 Internet is almost ossified Lots of band-aids and makeshift solutions (e.g. overlays) A new architecture (aka clean-slate) is needed Hard to come up with a one-size-fits-all architecture Almost impossible to predict what future might unleash Why not create an all-sizes-fit-into-one instead! Open and expandable architecture Testbed for future networking architectures and protocols
5
Network virtualization applications
Related Concepts 5 Virtual Private Networks (VPN) Virtual network connecting distributed sites Not customizable enough Active and Programmable Networks Customized network functionalities Programmable interfaces and active codes Overlay Networks Application layer virtual networks Not flexible enough
6
Business Model business model Players Relationships
7 Business Model Players Relationships Infrastructure Providers (InPs) Manage underlying physical networks Service Providers (SPs) Create and manage virtual networks Deploy customized end-to-end services End Users Buy and use services from different service providers Brokers Mediators/Arbiters End User SLA Service Provider Broker EIA NPA Infrastructure Provider SIA IIA
7
architecture 8 Architecture
8
Design Principles Design principles Hierarchy of Roles …
Concurrence of multiple heterogeneous virtual networks Introduces diversity Recursion of virtual networks Opens the door for network virtualization economics Inheritance of architectural attributes Promotes value-addition Revisitation of virtual nodes Simplifies network operation and management Hierarchy of Roles Service Provider N Infrastructure Virtual Network N Provider N+1 … Service Provider 1 Infrastructure Virtual Network 1 Provider 2 Service Provider 0 Infrastructure Virtual Network 0 Provider 1 Infrastructure Provider 0
9
Design goals Design Goals (1) Flexibility Manageability
10 Flexibility Service providers can choose arbitrary network topology, routing and forwarding functionalities, customized control and data planes No need for co-ordination with others IPv6 fiasco should never happen again Manageability Clear separation of policy from mechanism Defined accountability of infrastructure and service providers Modular management
10
Design goals Design Goals (2) Scalability
11 Scalability Maximize the number of co-existing virtual networks Increase resource utilization and amortize CAPEX and OPEX Security, Privacy, and Isolation Complete isolation between virtual networks Logical and resource Isolate faults, bugs, and misconfigurations Secured and private Design Goals (2)
11
Design goals Design Goals (3) Programmability Heterogeneity
12 Programmability Of network elements e.g. routers Answer “How much” and “how” Easy and effective without being vulnerable to threats Heterogeneity Networking technologies Optical, sensor, wireless etc. Virtual networks
12
Design goals Design Goals (4) Experimental and Deployment Facility
13 Experimental and Deployment Facility PlanetLab, GENI, VINI Directly deploy services in real world from the testing phase Legacy Support Consider the existing Internet as a member of the collection of multiple virtual Internets Very important to keep all concerned parties satisfied
13
Network virtualization summary
14 Definition (Sort of) Network virtualization is a networking environment that allows multiple service providers to dynamically compose multiple heterogeneous virtual networks that co-exist together in isolation from each other, and to deploy customized end-to-end services on-the-fly as well as manage them on those virtual networks for the end-users by effectively sharing and utilizing underlying network resources leased from multiple infrastructure providers.
14
Traditional router routing processor
Control plane and data plane embed in a blackbox designed by the vendor forwarding tables computed, pushed to input ports routing processor routing, management control plane (software) forwarding data plane (hardware) high-seed switching fabric router input ports router output ports
15
Software-Defined Networking
Decouple control plane from data plane Benefits: Fast innovation Global optimization Dynamic reconfiguration Flexibility Programmable network
16
OpenFlow OpenFlow – A representative SDN open protocol Secure channel
Flow table
17
Flow table Header fields Activity counter Actions
Header values for matching against packets Activity counter Update for matching packets Actions Actions taken on matching packets
18
Matching A packet matches a flow table entry if the values in the header fields used for the lookup (as defined above) match those defined in the flow table. If matching, perform the action and update the counter; If not matching, send the packet to the controller
19
Secure channel OpenFlow protocol Three types of messages:
Through this interface, the controller configures and manages the switch, receives events from the switch, and send packets out the switch. Three types of messages: Controller to switch Asynchronous Symmetric
20
Controller to switch Configuration Read state Modify state Send packet
The controller is able to set and query configuration parameters in the switch. Read state Read-State messages are used by the controller to collect statistics from the switchs flow-tables, ports and the individual flow entries. Modify state Their primary purpose is to add/delete and modify flows in the flow tables and to set switch port properties. Send packet These are used by the controller to send packets out of a specified port on the switch.
21
Asynchronous Packet-in Flow-removed Port-status Error
For all packets that do not have a matching flow entry, a packet-in event is sent to the controller (or if a packet matches an entry with a “send to controller" action). Fraction of packet header – sufficient buffer on switch Full packet – less buffer on switch Flow-removed Idle timeout – flow entries that lack activity Hard timeout – fixed lifetime of flow entry Port-status Error
22
Symmetric Hello Hello messages are exchanged between the switch and controller upon connection startup. Echo They can be used to indicate the latency, bandwidth, and/or liveness of a controller- switch connection. Can be used for efficient fault detection.
23
Network Functions Virtualisation Approach
Independent Software Vendors Message Router Session Border Controller WAN Acceleration CDN Carrier Grade NAT DPI Orchestrated, automatic remote install Tester/QoE monitor Firewall hypervisors Generic High Volume Ethernet Switches Generic High Volume Servers Generic High Volume Storage SGSN/GGSN The Classical network appliance approach uses a proprietary chassis for every new service, some typical examples are shown above left. Although inside the appliances use similar chips they are packaged very differently outside. This creates complexity through the entire life cycle of network services e.g. Design, procurement, test, deployment, configuration, repair, maintenance, replacement, end-of-life, removal. There is no economies of scale (some boxes may only sell in the thousands as opposed to 9 Million IT servers every year). The need for start-ups to develop new hardware, including getting it NEBs and ETSI qualified, is a significant cost and deterrent to enter the telecoms market. The Network Virtualisation (NV) approach replaces physical network appliances with software virtual appliances running on commodity IT servers. Using open IT techniques allows a competitive and innovative ecosystem to exploit the many x86 and Linux programmers in the world. The Virtual Appliances can be installed on IT servers using orchestration software, this will automatically and remotely install software, driven either by traffic demands or customer orders. To complement the standard high volume IT servers we will also use standard high volume storage and Ethernet switches. The standard high volume Ethernet switches will use merchant silicon which spreads the cost of switching ASICs over the widest possible Enterprise & Carrier market. BRAS Radio Network Controller PE Router Classical Network Appliance Approach 23
24
3 Complementary but Independent Networking Developments
Open Innovation Network Functions Virtualisation Software Defined Networks Creates operational flexibility Reduces Reduces CapEx, OpEx, space & power delivery time consumption Creates control abstractions to foster innovation. Creates competitive supply of innovative applications by third parties
25
Software defined multilayer networks
SDN controller L4-L7 SDN controller Extended Open Protocol Standardized Open Protocol Metro 1 WAN Data Center Metro 2 Optical Transport
26
Virtual network/infrastructure mapping
Virtual Infrastructures (VIs) SDN Controller Comp. Hypervisor Netw. Hypervisor Standardized Open Protocol Physical substrate
27
Problem statement Given: Find: Objective: A physical substrate;
A set of virtual requests; A set of modulation formats; Find: Mapping of virtual nodes over physical nodes; Mapping of virtual links over physical links; Routing, modulation and spectrum allocation of virtual links Objective: Maximize the number of accepted requests b a c d e 1 2 6 3 5 4 60 40 50 25 35 40G 10G 100G 75 65 30 10 60G 50G 25G 30G 140G 120G Virtual request Physical substrate Fig. 14 Virtual Infrastructure Mapping
28
Computing followed by Network Load Balance (CNLB)
Two-step solution: Computing load balance first Policy: map the virtual node onto the physical node with maximum available computing resources; Network load balance second Policy: calculate k-shortest path, map the virtual link to the physical route with the highest link mapping probability; 60 b 40 40G a 10G 100G c Virtual request 50 35 75 b Physical substrate 1 2 60G 30G 60G 10G 6 140G 3 50G c 30 25G 65 5 4 120G 40G a 10 60 Fig. 16 The CNLB Algorithm
29
Network followed by Computing Load Balance (NCLB)
Two-step solution Network load balance first Policy: check all the possible combinations of node pairs and select the physical link with the maximum link mapping probability; Computing load balance second Policy: after virtual link is mapped, map the attached two end nodes with computing load balance; 60 b 40 40G a 10G 100G c Virtual request 50 35 75 1 2 c Physical substrate 60G 30G 60G 10G 140G 6 3 50G 30 25G 65 5 4 120G 40G a 10 60 Fig. 17 The NCLB Algorithm
30
Upgrade-aware VI Mapping (U-VIM)
Basic scenario: VI upgrade in size or capacity, e.g., topology-varying General scenario: VI upgrade and physical substrate (PS) upgrade coexist Predict and Plan ahead: Provision UVT along with IVT and pre-reserve the demanded resources during the operation period Say this before S-VIM. Fig. 19 (a) VI upgrade; (b) PS upgrade
31
Survivable Virtual Infrastructure Mapping (S-VIM)
Given: A physical substrate, with computing resources at each physical node and spectrum resources at each physical link A set of virtual infrastructure request, with computing demand at each virtual node and networking demand at each virtual link Find: Mapping of primary and backup virtual nodes over physical nodes Routing, modulation selection, and spectrum allocation of primary and backup virtual links over physical links Objective: The number of mapped survivable virtual infrastructures are maximized. b a c d e 1 2 6 3 5 4 60 40 50 25 35 40G 10G 100G 75 65 105 95 80 60G 50G 35G 30G 160G 140G 120G Virtual request Physical substrate a’ I suggest you to first show a couple of slides on the Upgrade-aware VIM, to establish link with the previous topic on Predicitive Traffic Grooming. Doesn’t show any results. Then show the problem def of the S-VIM (this slide only) – remove the bottom Assumption and don’t show any results – just say you have publishe dpapers (add refs), c’ b’ d’ e’ Fig. 22 Survivable Virtual Infrastructure Mapping
32
S-VIM with shared protection
Virtual link protection: consider each virtual link as an individual component, and if two virtual link do not share any physical link, then their backup path can be shared. Virtual topology protection: construct a reliable virtual topology in the virtualization layer and then map the reliable virtual topology onto the physical substrate.
33
Enhanced protection to meet availability
Enhanced protection by iteratively adding redundancy: in order to meet the availability requirement, maybe one backup is not enough, so need to iteratively add more redundancy to improve the availability.
34
Intent-based network Standardize the north bound interface:
Hiding complexity from network application programmers and users Tell me the network characteristics you need And I will provide a network service to satisfy your requirement A LaTex version of setting up network service
35
Intent-based network Intent: “what”, not “how”
Intent as the “universal language” Intent is invariant Intent is portable Intent is compose-able Intent is scale-able Intent Brings Context
36
Joint Topology Design and Mapping (JTDM)
Telecom clouds (TCs): Computing resources Network function Single function TC (e.g., TC:D) Multi function TC (e.g., TC:A) Service function chains (SFCs): Virtual nodes (VNs) Computing power requirement Network function requirement Virtual links Bandwidth requirement Linear or irregular mesh JTDM VN combining SFC mapping
37
Challenges in JTDM VN combining SFC mapping Save bandwidth
Joint Optimization SFC mapping Network function requirement 80G 60G 2 80G 2 60G End Point 1 50G 4 End Point End Point 1 4 50G End Point 3 3 60G 50G Objective: min TBC 60G 50G B D B D f1 f2,f3 f2,f3 f1 A F A F f1,f2 f4 f1,f2 f4 C E C E f1, f3 f2 f2 f1, f3 (a) No VN combining (b) VN combining Total bandwidth consumption = 300 Gbps Total bandwidth consumption = 450 Gbps
Similar presentations
