1. CSNS 831 - SOFTWARE DEFINED NETWORKS K. PALANIVEL Dept. of Computer Science / Computer Centre Pondicherry University, Puducherry - 605014 Lecture 2: How does SDN work

2. Topics to be Discussed SDN Origins and Evolution Introduction Why SDN? Centralized and Distributed Control and Data Planes The Genesis of SDN Department of Computer Science, Pondicherry University

3. INTRODUCTION

4. Thanks to the advances in today’s off-the-shelf hardware or Whitebox networking, developer tools and standards, a seismic technology shift in networking to software can finally take place.

5. Problems in the Existing Networks Programmable networks Intelligence and control centralized Network interaction via APIs Vendor-neutral architectures There are a number of switch and router vendors that have announced their support of OpenFlow, including Cisco, Juniper, Big Switch Networks, Brocade, Arista, Extreme Networks, IBM, Dell, NoviFlow, HP, NEC, among others.CiscoJuniperBig Switch NetworksBrocadeAristaExtreme NetworksIBMDellNoviFlowHPNEC Department of Computer Science, Pondicherry University

6. SDN separates the control of the network from the hardware. It uses software applications to program your network intelligently through centralized control. This means the underlying hardware and associated technologies are still there, but they are programmed centrally. As a result, you can consistently and holistically manage your entire network with utmost flexibility and speed. Department of Computer Science, Pondicherry University Problems in the Existing Networks

7. Definition “an architecture that aims to make networks agile and flexible”. “The physical separation of the network control plane from the forwarding plane, and where a control plane controls several devices” “a network architecture approach that enables the network to be intelligently and centrally controlled, or ‘programmed,’ using software applications.” “An approach to cloud computing that facilitates network management and enables programmatically efficient network configuration in order to improve network performance and monitoring”. (Wikipedia) Goal: is to improve network control by enabling enterprises and service providers to respond quickly to changing business requirements. Department of Computer Science, Pondicherry University

8. What is Software-defined Networking? Software-defined networking (SDN) is designed to make a network flexible and agile. It lets you design, build, and manage networks, separating the control and forwarding planes. As a result, the control plane is directly programmable, and it abstracts the underlying infrastructure for applications and network services. Network intelligence is logically centralized through programmable SDN controllers. Implemented in software, these controllers maintain a coherent view of the network domain. To applications and policy engines, SDN looks like a single logical switch. Department of Computer Science, Pondicherry University

9. What is software-defined networking?... Software Defined Networking (SDN) developed out of the need to automate, scale and optimize networking for applications that may be provided either via an enterprise datacenter, a Virtual Private Cloud (VPC), or as-a-service (public cloud). SDN as a centralized approach to the management of network infrastructure. SDN provides a number of important benefits for network and IT operators through controller-enabled, network visibility and automation including: Ability to programmatically automate network configurations, increasing scalability and reliability Increased flexibility and agility for changing the network operation to enable an application or address a task. Centralized visibility of the network topology, network elements and their operation across the network infrastructure. Department of Computer Science, Pondicherry University

10. What does SDN consist of? Software-defined networking (SDN) offers a centralized, programmable network that consists of an SDN controller, southbound APIs, and northbound APIs. SDN controllers are the brains of the network, offering a centralized view of the overall network. Southbound APIs relay information to the switches and routers in network. Northbound APIs communicate with the applications and deploy services. Department of Computer Science, Pondicherry University

11. SDN Model & Architecture Department of Computer Science, Pondicherry University

12. Benefits of SDN Fast and automated application deployment. On-demand app delivery and mobility at scale. Greater resource flexibility and utilization. Reduction of IT costs by enhancing the benefits of data center virtualization. Department of Computer Science, Pondicherry University

13. Relationship between SDN and intent-based networking? Software-defined networks have automated the process of network management. Intent-based networking adds context, learning, and assurance capabilities by tightly coupling policy with intent. SDN can be a building block of intent-based networking (the network what you want (your intent) rather than exactly what to do, on a technical level, to accomplish your goal.)intent-based networking Department of Computer Science, Pondicherry University

14. Complementary Approaches NFV, Network Virtualization, and White Box Networking (bare metal switching) are all complementary approaches. They each offer a new way to design deploy and manage the network and its services: SDN: separates the network’s control (brains) and forwarding (muscle) planes and provides a centralized view of the distributed network for more efficient orchestration and automation of network services. NFV (Network Functions Virtualization): focuses on optimizing the network services themselves. NFV decouples the network functions, such as DNS, caching, etc., from proprietary hardware appliances, so they can run in software to accelerate service innovation and provisioning, particularly within service provider environments. Department of Computer Science, Pondicherry University

15. Complementary Approaches NV (Network Virtualization): ensures the network can integrate with and support the demands of virtualized architectures, particularly those with multi-tenancy requirements. White Box: uses network devices, such as switches and routers, that as based on “generic” merchant silicon networking chipset available for anyone to buy, as opposed to proprietary silicon chips designed by and for a single networking vendor. Department of Computer Science, Pondicherry University

16. Relationship between SDN and NFV Network function virtualization (NFV) uses hypervisor and cloud- computing technology for network automation and orchestration. NFV works best in the context of network services (OSI Level 4 and up) that require heavy compute power, with low-to-medium bandwidth throughput. SDN converges the management of network and application services into centralized, extensible orchestration platforms. SDN is optimal for high-throughput network forwarding (OSI Levels 0-3) where bandwidth-intensive workloads need significant traffic management.high-throughput network forwarding Department of Computer Science, Pondicherry University

17. SDN Projects ONOS - Open Network Operating System (ONOS), a Linux Foundation project, is a software-defined networking OS for service providers that has scalability, high availability, high performance and abstractions to create apps and services. OPENCONTRAIL - OpenContrail is Juniper Networks’ open source network virtualization platform for the cloud. It provides all the necessary components for network virtualization: SDN controller, virtual router, analytics engine, and published northbound APIs. Its REST API configures and gathers operational and analytics data from the system. OPENDAYLIGHT - OpenDaylight, an OpenDaylight Foundation project at The Linux Foundation, is a programmable, software-defined networking platform for service providers and enterprises. Based on a microservices architecture, it enables network services across a spectrum of hardware in multivendor environments. Department of Computer Science, Pondicherry University

18. OPEN VSWITCH - Open vSwitch, a Linux Foundation project, is a production-quality, multilayer virtual switch. It’s designed for massive network automation through programmatic extension, while still supporting standard management interfaces and protocols including NetFlow, sFlow, IPFIX, RSPAN, CLI, LACP, and 802.1ag. It supports distribution across multiple physical servers similar to VMware’s vNetwork distributed vswitch or Cisco’s Nexus 1000V. OPNFV - Open Platform for Network Functions Virtualization (OPNFV), a Linux Foundation project, is a reference NFV platform for enterprise and service provider networks. It brings together upstream components across compute, storage and network virtualization in order create an end-to-end platform for NFV applications. SDN Projects Department of Computer Science, Pondicherry University

19. TRADIONAL NETWORK CONFIGURATION Department of Computer Science, Pondicherry University

20. Traditional Network Configuration The traditional approach to networking is characterized by two main factors: 1.Network functionality is mainly implemented in a dedicated appliance. In this case, ‘dedicated appliance’ refers to one or multiple switches, routers and/or application delivery controllers. 2.Most functionality within this appliance is implemented in dedicated hardware. An Application Specific Integrated Circuit (or: ASIC) is often used for this purpose. Department of Computer Science, Pondicherry University

21. How routers route packets from the source to the destination? Department of Computer Science, Pondicherry University

22. Router Architecture Department of Computer Science, Pondicherry University

23. How switches route packets from the source to the destination? Department of Computer Science, Pondicherry University

24. Switch Architecture Department of Computer Science, Pondicherry University

25. Network Architecture (Planes) Department of Computer Science, Pondicherry University

26. Network Architecture (Protocols) Department of Computer Science, Pondicherry University

27. From Traditional Network to SDN Department of Computer Science, Pondicherry University

28. From Traditional Network to SDN Department of Computer Science, Pondicherry University

29. SDN Architecture – View 1 Department of Computer Science, Pondicherry University

30. SDN Architecture – View 2 Department of Computer Science, Pondicherry University

31. Limitations Organizations are increasingly confronted with the limitations that accompany this hardware-centric approach, such as: Traditional configuration is time-consuming and error-prone Multi-vendor environments require a high level of expertise Traditional architectures complicate network segmentation Embracing change: Software Defined Networking? Department of Computer Science, Pondicherry University

32. Three Planes Forwarding Plane - Moves packets from input to output Control Plane - Determines how packets should be forwarded Management Plane - Methods of configuring the control plane (CLI, SNMP, etc.) Department of Computer Science, Pondicherry University

33. Limitations This concerns the two network device planes, i.e.: The plane that determines where to send traffic (control plane) The plane that executes these decisions and forwards traffic (data plane) Department of Computer Science, Pondicherry University

34. Software Layer Decoupling these two planes involves leaving the data plane with network hardware and moving the control plane into a software layer. By abstracting the network from the hardware, policies no longer have to be executed on the hardware itself. Instead, the use of a centralized software application functioning as the control plane makes network virtualization possible. This process is similar to server virtualization Server virtualization: The process of creating various VMs (virtual machines) and decoupling them from physical servers. Network virtualization: The process of creating virtual networks which are decoupled from physical network components. Department of Computer Science, Pondicherry University

35. SDN BENEFITS Department of Computer Science, Pondicherry University

36. SDN Business Benefits Software Defined Networking is expected to have several business benefits, including: More configuration accuracy, consistency and flexibility: As described earlier, traditional networking requires configurations to be executed on a manual, deviceby- device basis. A key characteristic of the SDN approach, is automating this process, enabling an administrator to manage the entire network as if it were a single device (see figure 2). In addition to increasing configuration accuracy and consistency, this method also boosts a networks’ responsiveness. In case network conditions change, an administrator can adjust existing configurations much quicker. Data flow optimization: A second expected business benefit of the SDN approach, is the optimization of data flows. Instead of having a single path from the source of communication flow to its destination, a SDN controller is able to identify multiple paths per flow. Furthermore, this approach allows the flow’s traffic to be split across multiple nodes. Network performance and scalability is enhanced by optimizing the network path for a particular data flow based on the source and destination nodes1. Department of Computer Science, Pondicherry University

37. Why SDN? Enable Innovation: enabling organizations to create new types of applications, services and business models Offer New Services: Create new revenue generating services Reduce CapEx: allowing network functions to run on off-the-shelf hardware Reduce OpEX: supporting automation and algorithm control through increased programmability of network elements to make it simple to design, deploy, manage and scale networks Deliver Agility and Flexibility: helping organizations rapidly deploy new applications, services and infrastructure to quickly meet their changing requirements Department of Computer Science, Pondicherry University

38. Why SDN? Programmable networks Intelligence and control centralized Network interaction via APIs Vendor-neutral architectures Department of Computer Science, Pondicherry University

39. Why SDN?: Programmable networks Historically, your network was only as good as the hardware that controlled it. SDN changes all that and provides easy customizations, even down to the individual customer level. With hardware decoupled from software, you can introduce innovative, differentiated new services quickly—something previously unimagined with the constraints of closed and proprietary platforms. Department of Computer Science, Pondicherry University

40. Why SDN?: Intelligence and Control Centralized Bandwidth management, restoration, security, and policies have been a thorn in the side of every network operator. Now, with those functions centrally controlled by an SDN controller that is highly intelligent and optimized, you now have a holistic view of the network. That is an asset to your operations, not a legacy liability. With network control centralized, your network resources can be controlled and managed in a coordinated way to delivery services end to end. And devices now operate with awareness of the conditions of the network overall. Department of Computer Science, Pondicherry University

41. Why SDN?: Network Interaction via APIs Your static physical hardware and network connections are a thing of the past with SDN. Services and applications are no longer tied to network hardware and connections. Instead, your applications connect over the network infrastructure flexibly, with the use of APIs, between OSS/BSS, orchestration, and assurance systems. Department of Computer Science, Pondicherry University

42. Why SDN?: Vendor-neutral architectures SDN enables an open approach that is vendor-neutral and supports a large assortment of applications. Cloud orchestration, SaaS, and business-critical networked apps are just a few of the possibilities SDN enables. With SDN, intelligent network services and applications run in a common IT environment that can control the hardware and associated technologies of countless vendors. Department of Computer Science, Pondicherry University

43. Three Planes Forwarding Plane - Moves packets from input to output Control Plane - Determines how packets should be forwarded Management Plane - Methods of configuring the control plane (CLI, SNMP, etc.) Department of Computer Science, Pondicherry University

44. Control Planes Control Plane - The control plane and management plane serve the data plane, which bears the traffic that the network exists to carry. Makes decisions about where traffic is sent Control plane packets are destined to or locally originated by the router itself The control plane functions include the system configuration, management, and exchange of routing table information The route controller exchanges the topology information with other routers and constructs a routing table based on a routing protocol, for example, RIP, OSPF or BGP Control plane packets are processed by the router to update the routing table information. It is the Signalling of the network Since the control functions are not performed on each arriving individual packet, they do not have a strict speed constraint and are less time-critical Department of Computer Science, Pondicherry University

45. Example 1 The protocol or application itself doesn’t really determine whether the traffic is control, management, or data plane, but more importantly how the router processes it. Consider a 3 router topology with routers R1, R2 and R3. Lets say a Telnet session is established from R1 to R3. On both of these routers the packets need to be handled by the control/management plane. However from R2′s perspective this is just data plane traffic that is transiting between its links. Department of Computer Science, Pondicherry University

46. Example 2 Control Plane => Learning what we will do Our planning stage, which includes learning which paths the buses will take, is similar to the control plane in the network. We haven’t picked up people yet, nor have we dropped them off, but we do know the paths and stops due to our plan. The control plane is primarily about the learning of routes. Data Plane => Actually moving the packets based on what we learned. The data plane is the actual movement of the customers data packets over the transit path we learned in the control plane stage. Department of Computer Science, Pondicherry University

47. Data Plane Data Plane : The data plane (sometimes known as the user plane, forwarding plane, carrier plane or bearer plane) is the part of a network that carries user traffic. The data plane, the control plane and the management plane are the three basic components of a telecommunications architecture.architecture Also known as Forwarding Plane Forwards traffic to the next hop along the path to the selected destination network according to control plane logic Data plane packets go through the router The routers/switches use what the control plane built to dispose of incoming and outgoing frames and packets Department of Computer Science, Pondicherry University

48. The Control Plane, Data Plane and Forwarding Plane in Networks Department of Computer Science, Pondicherry University

49. Planes of Operation The Control Plane, Data Plane and Forwarding Plane in Networks is the heart core DNA (Database Network Associates) in today’s networking hardware to move IP packets from A to Z. The Management plane is another vital component but also widely excepted as user to hardware interaction. These planes of operation are the building blocks of the layered architecture that networks have evolved to today. Department of Computer Science, Pondicherry University

50. Control Plane The control plane is the component to a router that focuses on how that one individual box interacts with its neighbors with state exchange. The Routing Information (data)Base (RIB) and Label Information Base (LIB) are processed in software and used to populate FIB(forwarding information base) and the LFIB. Vendors can implement these in different fashions on how those tables are partitioned between multiple routing instances. For example, a router has a BGP and OSPF adjacencies, those routing protocols have different algorithms to determine what a chosen path to a network would be. Building the topology or global view as that particular router sees it from its point of view. That is fairly important to recognize that its “global view” is from its perspective of either the IGP or EGP. Department of Computer Science, Pondicherry University

51. The Control plane The Control plane feeds the forwarding/data plane with what it needs to create its forwarding tables and updates topology changes as they occur. Those are pretty low even in large networks single to at most I would speculate double digit per second changes. This is the reason the control plane can often be thought of as the “slow path” in legacy route once switch many packet switching architectures. Department of Computer Science, Pondicherry University

52. The Control plane A list of functions performed in traditional routing engines/route processors are the following: Allocates resources to the forwarding engine/plane. Routing state ARP handling is always processed by general purpose processor located in the routing engine. Security functions to secure the control plane access. Telnet, ssh, AAA etc. Establishes and maintains management sessions, such as Telnet connections Routing state to neighboring network elements. Vendor and platform specific stacking, clustering, pairing etc. Department of Computer Science, Pondicherry University

53. The Planes separated and typical packet logic into these information bases. Department of Computer Science, Pondicherry University

54. Break the control and data forwarding plane up Department of Computer Science, Pondicherry University

55. Slow and Fast path in a software vSwitch (for those that have a Control Plane). Department of Computer Science, Pondicherry University

56. Centralized or Distributed Forwarding data planes Forwarding data planes typically come either centralized or distributed. This means the forwaring engine is either centrally located across the ethernet fabric/crossbar or pushed all the way to the edge. The more performance required the more that distributed forwarding is pushed to the edge. How we design networks is merely a macro of what is happening on the board.macro Department of Computer Science, Pondicherry University

57. Centralized or Distributed Forwarding data planes Department of Computer Science, Pondicherry University

58. Management Plane Telnet, SNMP, SSH, XML, stone tablet and chisel, thats some of how we orchestrate and operate networks today, well except for the wireless market.wireless It is pretty simple today. I think that says a lot about the industry right now for that matter. If we get anything out of the next few years it will be simplification of network operations and management from abstraction, orchestration and/or automation. There have been some interesting thoughts and IETF draft submissions lately around the operational plane. Department of Computer Science, Pondicherry University

59. Thank You Department of Computer Science, Pondicherry University