What is software-defined networking (SDN) and the architecture

SDN (Software Defined Networking) is the latest buzzword in IT, getting more popular every year. In simple word we can say that the goal of Software-Defines Networking (SDN) is to enable cloud computing and network engineers and administrators to respond quickly to changing business requirements via a centralized control console.

                                                   Or

Software-defined networking (SDN) is an architecture that aims to make networks agile and flexible. The goal of SDN is to improve network control by enabling enterprises and service providers to respond quickly to changing business requirements.In a software-defined network, a network engineer or administrator can shape traffic from a centralized control console without having to touch individual switches in the network. The centralized SDN controller directs the switches to deliver network services wherever they're needed, regardless of the specific connections between a server and devices.

SDN and SD-WAN

While on the other hand, SDN is primarily focused internally, within the LAN (locally) or within the Service Provider's core network. ... SDN is focused on the internal network, be it the LAN or the core service provider network. While SD-WAN is focused on enabling connections between networks and users over the WAN.Jun 8, 2017

What is SDN, why is it becoming popular?

To answer these questions, we will have to take a closer look at “traditional” networking first. We will discuss the current “limitations” of traditional networking, I will explain what SDN is and how SDN is supposed to solve the “problems” that traditional networking has.

Traditional Networking

Networking has always been very traditional. We have specific network devices like routers, switches, and firewalls that are used for specific tasks.

A network device, for example, a router has different functions that it has to perform. Think for a moment about some of the things that a router has to do in order to forward an IP packet:

• It has to check the destination IP address in the routing table in order to figure out where to forward the IP packet to.
• Routing protocols like OSPF, EIGRP or BGP are required to learn networks that are installed in the routing table.
• It has to use ARP to figure out the destination MAC address of the next hop or destination and change the destination MAC address in the Ethernet frame.
• The TTL (Time to Live) in the IP packet has to be decreased by 1 and the IP header checksum has to be recalculated.
• The Ethernet frame checksum has to be recalculated.

All these different tasks are separated by different planes. There are three planes:

• control plane
• data plane
• management plane

Let’s take a look at the difference between these three planes…

Control Plane

The control plane is responsible for exchanging routing information, building the ARP table, etc. Here are some tasks that are performed by the control plane:

• Learning MAC addresses to build a switch MAC address table.
• Running STP to create a loop-free topology.
• Building ARP tables.
• Running routing protocols like OSPF, EIGRP, and BGP and building the routing table.

Data Plane

The data plane is responsible for forwarding traffic. It relies on the information that the control plane supplies. Here are some tasks that the data plane takes care of:

• Encapsulate and de-encapsulate packets.
• Adding or removing headers like the 802.1Q header.
• Matching MAC addresses for forwarding.
• Matching IP destinations in the routing table.
• Change source and destination addresses when using NAT.
• Dropping traffic because of access-lists.

The tasks of the data plane have to be performed as fast as possible which is why the forwarding of traffic is performed by specialized hardware like ASICs and TCAM tables.

Management Plane

The management plane is used for access and management of our network devices. For example, accessing our device through telnet, SSH or the console port.

When discussing SDN, the control and data plane are the most important to keep in mind. Here’s an illustration of the control and data plane to help you visualize the different planes:

Above you can see the control plane where we use routing protocols like OSPF and EIGRP and some static routing. The best routes are installed in the routing table. Another table that the router has to build is the ARP table.

Information from the routing and ARP table is then used to build the forwarding table. When the router receives an IP packet, it will be able to forward it quickly since the forwarding table has already been built.

Limitations of traditional networking

Everything I described above is the way we have done things for the last ~30 years so it’s not like there is something “wrong” with traditional networking. However, nowadays there are some business challenges that ask for different solutions.

Traditional networking uses a distributed model for the control plane. Protocols like ARP, STP, OSPF, EIGRP, BGP and other run separately on each network device. These network devices communicate with each other but there is no central device that has an overview or that controls the entire network.

One exception here (for those that are familiar with wireless networking) are the wireless LAN controllers (WLC). When you configure a wireless network, you configure everything on the WLC which controls and configures the access points. We don’t have to configure each access point separately anymore, it’s all done by the WLC.

So, this is the reason SDN came into the picture which is nothing but a process to move away from traditional network architecture, in which individual network devices make traffic decisions based on their configured routing tables, and same is the reason behind emerging the new technology SD-WAN that is implementing the same into WAN as well, we can quickly have a look into that as well before moving further.

As we too have problem with traditional WAN connections like cost (private WAN connections like MPLS are way more expensive than regular Internet connections), time to deploy, QoS, SLA and packet loss.

The way we use our WAN has also changed throughout the years. Most organizations had an HQ, remote users, and perhaps some branch offices. Branch offices were connected to the HQ with private WAN or VPNs over the Internet. Remote users used remote VPN over the Internet to connect.

Nowadays, organizations also run their own applications in the cloud instead of on-premises, and they use applications like Office 365 or Gsuite. Our traffic patterns look different now:

What about network management? Each router has its own control plane, and we use the CLI to manually create our router configurations “box-by-box”. This is time-consuming and prone to errors. We can use network automation tools to make our lives easier, but the control plane remains decentralized, so to resolve this issue SD-WAN came into the picture.

SDN architecture

A typical representation of SDN architecture comprises three layers: the application layer, the control layer and the infrastructure layer.

SDN architecture separates the network into three distinguishable layers, connected through northbound and southbound APIs.

The application layer, not surprisingly, contains the typical network applications or functions organizations use, which can include intrusion detection systems, load balancing or firewalls. Where a traditional network would use a specialized appliance, such as a firewall or load balancer, a software-defined network replaces the appliance with an application that uses the controller to manage data plane behavior.

The control layer represents the centralized SDN controller software that acts as the brain of the software-defined network and is built with the network automation. This controller resides on a server and manages policies and the flow of traffic throughout the network.We create the network policy globally and push them to all the routers from central location .You create a QoS policy and push them to all the 500 routers from a central location.

The infrastructure layer is made up of the physical switches in the network.

These three layers communicate using respective northbound and southbound application programming interfaces (APIs). For example, applications talk to the controller through its northbound interface, while the controller and switches communicate using southbound interfaces, such as OpenFlow (OpenFlow is a protocol that allows a server to tell network switches where to send packets. In a conventional network, each switch has proprietary software that tells it what to do. With OpenFlow, the packet-moving decisions are centralized, so that the network can be programmed independently of the individual switches and data center gear)although other protocols exist.

How SDN works

SDN encompasses several types of technologies, including functional separation, network virtualization and automation through programmability.

Originally, SDN technology focused solely on separation of the network control plane from the data plane. While the control plane makes decisions about how packets should flow through the network, the data plane actually moves packets from place to place.

In a classic SDN scenario, a packet arrives at a network switch, and rules built into the switch's proprietary firmware tell the switch where to forward the packet. These packet-handling rules are sent to the switch from the centralized controller.

The switch also known as a data plane device queries the controller for guidance as needed, and it provides the controller with information about traffic it handles. The switch sends every packet going to the same destination along the same path and treats all the packets the exact same way.

Software-defined networking uses an operation mode that is sometimes called adaptive or dynamic, in which a switch issues a route request to a controller for a packet that does not have a specific route. This process is separate from adaptive routing, which issues route requests through routers and algorithms based on the network topology, not through a controller.

Benefits of SDN

With SDN, an administrator can change any network switch's rules when necessary -- prioritizing, reprioritizing or even blocking specific types of packets with a granular level of control and security. This is especially helpful in a cloud computing multi-tenant architecture, because it enables the administrator to manage traffic loads in a flexible and more efficient manner. Essentially, this enables the administrator to use less expensive commodity switches and have more control over network traffic flow than ever before.

Other benefits of SDN are network management and end-to-end visibility. A network administrator need only deal with one centralized controller to distribute policies to the connected switches, instead of configuring multiple individual devices. This capability is also a security advantage because the controller can monitor traffic and deploy security policies. If the controller deems traffic suspicious, for example, it can reroute or drop the packets.

SDN also virtualizes hardware and services that were previously carried out by dedicated hardware, resulting in the touted benefits of a reduced hardware footprint and lower operational costs.

Challenges with SDN

Security is both a benefit and a concern with SDN technology. The centralized SDN controller presents a single point of failure and, if targeted by an attacker, can prove detrimental to the network.

Ironically, another challenge with SDN is there's really no established definition of software-defined networking in the networking industry. Different vendors offer various approaches to SDN, ranging from hardware-centric models and virtualization platforms to hyper-converged networking designs and controllerless methods.