Network Function Virtualization in Software Defined Optical Transport Networks

Transcription

1 Network Function Virtualization in Software Defined Optical Transport Networks Yongli Zhao State Key Laboratory of Information Photonics and Optical Communications, Beijing University of Posts and Telecommunications (BUPT), Beijing , China Tel : ; yonglizhao@bupt.edu.cn; Web: ipoc.bupt.edu.cn Group Meeting, June 17, 2016

2 Outline NFV enabled software defined multi-domain optical networks Procedure of call and return NFV enabled optical transport node architecture Parallel processing method for NFV enabled OTN nodes Conclusion Page 2

3 What is Network Functions Virtualization (NFV)? Network Functions Virtualization aims to transform the way that network operators architect networks by evolving standard IT virtualization technology to consolidate many network equipment types onto industry standard high volume servers, switches and storage, which could be located in Datacenters, Network Nodes and in the end user premises. Network Functions Virtualization Introductory White Paper Page 3

4 Benefits of Network Functions Virtualization Reduced equipment costs and reduced power consumption through consolidating equipment and exploiting the economies of scale of the IT industry. Increased velocity of Time to Market by minimizing the typical network operator cycle of innovation. Targeted service introduction based on geography or customer sets is possible. Services can be rapidly scaled up/down as required. Enabling a wide variety of eco-systems and encouraging openness. Optimizing network configuration and/or topology in near real time based on the actual traffic/mobility patterns and service demand.. Network Functions Virtualization Introductory White Paper Page 4

5 Relationship with Software Defined Networks (SDN) Network Functions Virtualization is highly complementary to Software Defined Networks (SDN), but not dependent on it. Network Functions Virtualization can be implemented without a SDN being required, although the two concepts and solutions can be combined and potentially greater value accrued. NFV can be achieved using non-sdn mechanisms. But approaches relying on the separation of the control and data forwarding planes as proposed by SDN can enhance the performance. NFV is able to support SDN by providing the infrastructure upon which the SDN software can be run. Furthermore, NFV aligns closely with the SDN objectives to use commodity servers and switches Network Functions Virtualization Introductory White Paper Page 5

6 Fields of Application and Use Cases Switching elements: BNG, CG-NAT, routers. Mobile network nodes: HLR/HSS, MME, SGSN, GGSN/PDN-GW, RNC, Node B, enode B. Functions contained in home routers and set top boxes to create virtualized home environments. Tunneling gateway elements: IPSec/SSL VPN gateways. Traffic analysis: DPI, QoE measurement. Service Assurance, SLA monitoring, Test and Diagnostics. NGN signaling: SBCs, IMS. Converged and network-wide functions: AAA servers, policy control and charging platforms. Application-level optimization: CDNs, Cache Servers, Load Balancers, Application Accelerators. Security functions: Firewalls, virus scanners, intrusion detection systems, spam protection. How about optical transport networks? Network Functions Virtualization Introductory White Paper Page 6

7 Optical Service Chaining for Network Function Virtualization NFV traffic patterns are expected to fluctuate substantially compared to the legacy network services, because s can be dynamically created and destroyed as needed. The configuration of matching rules can be complicated and error-prone as flow number goes large, and may lead to high operational complexity and cost. Ming Xia, et al., Optical Service Chaining for Network Function Virtualization Network, IEEE Communications Magazine, April Page 7

8 Optical Service Chaining for Network Function Virtualization It introduces an optical steering domain into the operator s data centers for NFV service chaining at a coarsegrained traffic level using wavelength switching. In the proposed architecture, a service chain may partly reside in the optical domain and partly reside in the packet domain. Ming Xia, et al., Optical Service Chaining for Network Function Virtualization Network, IEEE Communications Magazine, April Page 8

9 Transport NFV Architecture NFV concepts do not only apply to data plane functions (i.e., packet processing or forwarding), but also to control plane functions, such as path computation. First, they focus on the IT and network resources that are virtualized to provide the required VNFs. Second, they provide an example of VNF on top of the virtualized infrastructure, by proposing a path computation element (PCE) architecture to deploy a PCE by means of NFV. Ricard Vilalta, et al., Transport Network Function Virtualization, Journal of Lightware Technology, vol.33, no.8, Apr Page 9

10 Data plane Control plane APP plane NFV enabled software defined multi-domain optical networks Application Plane Controller Database TED Config Extended OpenFlow Plugin Control Board OpenFlow Gateway Vrouter APP APP APP APP APP APP APP (routing) NorthBound Component Configuration Orchestrator Database TED Extended OpenFlow Plugin Basic functions Connection Control Config (resource) Interface Extended OpenFlow Protocol Topology Manager NorthBound Call/return Call/return Component Configuration (grooming) Basic functions Connection Control Extended OpenFlow Protocol Call/return Controller Database TED Interface Topology Manager (strategy) Config Extended OpenFlow Plugin CVNI (management) NorthBound Component Configuration CVNI RESTful Basic functions Connection Control Interface Extended OpenFlow Protocol OpenFlow Topology Manager Control Board OpenFlow Gateway Vrouter Application (APP) plane contains various applications, such as bandwidth on demand (BoD), network on demand (NoD), and service scheduling expert (SSE). These service requests are initialed by this plane, and served by the network orchestrator through RESTful APIs interfaces. Control plane is composed of two kinds of controllers. Each domain owns one local controller, and the local domain controllers are managed by one super controller, which is named as network orchestrator. Developed based on Open Daylight controller, the orchestrator is responsible for communicating with controllers of different domains via control virtual network interface (CVNI), which is implemented based on the extended Openflow protocol. Based on the implementation mechanism of different function modules with software engineering, NFV is considered to be embedded into the control plane. Some function modules can be abstracted from the controller as, such as path computation, resource allocation, grooming algorithm, security strategy, and user management. Both orchestrator and local controllers can call the necessary functions from different s, and then the related returns the result. Data plane consists of all the transport equipment controlled by SDON controllers. In each domain, the transport nodes are equipped with specialized control boards to communicate with their domain controller via the domainspecified protocol, i.e, OpenFlow protocol. Yongli Zhao, et al., Network Function Virtualization in Software Defined Optical Transport Networks, OFC2016, Anaheim, CA, USA, Mar Page 10

11 Procedure of call and return n Controller Orchestrator Controller APP TCP Three-way Handshake TCP Three-way Handshake OpenFlow Handshake & Keepalive OpenFlow Handshake & Keepalive Port List Request & Reply Port List Request & Reply Sercice Request In Local Domain Service Request Service Request In Local Domain 1 Call 2 Call n Call 1 Return 2 Return n Return Service Reply In Local Domain Service Reply In Local Domain Service Reply Yongli Zhao, et al., Network Function Virtualization in Software Defined Optical Transport Networks, OFC2016, Anaheim, CA, USA, Mar Page 11

12 NFV enabled optical transport node architecture Interface card Interface card Interface card IF IF IF IF IF IF IF IF IF 10G 10G 10G 10G 10G 10G 1G 1G 1G (Forwarding) (Flow table) (Packet switching) Layer 2 processing... (Ether. Enc/Dec ) Data processing fabric (OTU Mux) (OTU De-mux) (k switching) Layer 1 processing... (OTU FEC ) OXC Tx Rx Tx Rx Tx Rx Tx Rx Tx Rx Tx Rx Transceiver Pool E W S N Line side Yongli Zhao, et al., Network Function Virtualization in Software Defined Optical Transport Networks, OFC2016, Anaheim, CA, USA, Mar Page 12

13 Data distributer NFV enabled optical transport node architecture Optical interface Optical interface Optical interface Electronic interface O/E/O O/E/O O/E/O OTN process unit OTN process unit O/E/O O/E/O O/E/O OTN process unit OTN process unit mana ger OCH OTN process unit OMS OTS OTN process unit Virtual machine cluster for Main optical path Main optical path Clock X86 platform Page 13

14 Data distributer Parallel processing method for NFV enabled OTN nodes T4 T3 T2 T1 T0 Service port 1 T4 T3 T2 T1 T0 Line port 1 T4 T3 T2 T1 T0 Line port 2 OTUk adap-1 OTUk adap-2 OTUk adap-3 OTUk adap-4 OTUk adap-5 OTUk adap-6 OTUk adap-1 OTUk adap-2 OTUk adap-3 OTUk adap-4 OTUk adap-5 OTUk adap-6 k Cross-1 k Cross-2 k Cross-3 k Cross-4 k Cross-5 k Cross-6 Input OEO 1. The distributer 2. k adap schedules the data to 4. OTUk adap modules in circle 3. k Cross OEO Page 14

15 Use case Optical Cross Connector processing 120Gbps 120Gbps! Optical Cross Connector BUS 20Gbps Load per X86 Platform X86 Platform lambda Wavelength device lambda WDM device Port 1 Port 3 Port 1 Port 3 Port 2 Port 2 Page 15

16 Use case Optical Cross Connector 20Gbps Load per Optical Cross Connector BUS 20Gbps Load per Optical Cross Connector 20Gbps Load per BUS X86 Platform BUS lambda X86 Platform WDM device lambda WDM device lambda X86 Platform WDM device Port 1 Port 1 Port 3 Port 3 Port 1 Port 3 Port 2 Port 2 Port 2 Step 1 Step 2 Step 3 Page 16

17 Conclusion This paper introduce NFV into software defined optical transport networks (SDON). NFV enabled software defined multi-domain optical networks is designed with procedure demonstration of call and return. NFV based control plane and transport nodes are designed to support NFV enabled SDON. Parallel processing method is proposed for NFV enabled OTN nodes. Page 17

18 Thank You! Beijing University of Posts and Telecommunications