A software-defined wide area network (SD-WAN) is a smart way to manage your network. It connects sites, branches, and data centres, using multiple connectivity options such as broadband, MPLS and LTE.
SD-WAN centralises the control of a network, the provisioning of new services and the ability to make application focused routing decisions based on software intelligence. It provides increased bandwidth or Internet connectivity at a lower cost, better control, much simpler network management and reduced overall OpEx.
Let’s start with the WAN in SD-WAN
A wide-area network, or WAN, connects different users at different sites to the applications they use. A WAN consists of physical infrastructure like routers, switches and transmission devices such as satellite and fibre.
If a business has multiple branches that require site-to-site connectivity, access to applications hosted in a central environment like a datacentre that sits at the head office and all branches require internet access.
To deliver site-to-site connectivity, access to central applications and internet access, data will flow from a branch over the link and into the central environment (head office) from there the data will either be transmitted to the other branches, access the resource hosted in the private environment or break out to the internet via a centralised point.
So, the WAN would look something like this:
Employees at Branch A, as well as at Branches B, C,and D, connect to these applications over the WAN.
The CRM that all branch employees use is located at Branch B. Employees at Branch A, C and D connect to this application over the WAN.
And all employees, at the head office and other branches, need to connect to the Internet via a centralised breakout point on the MPLS network.
WAN: robust but static
In order to optimise performance on a traditional WAN this needed to be done at a granular level, link by link, site by site. This meant paying attention to the network infrastructure and how it was configured.
The WAN could be considered as relatively static. Once it has been set up, it is difficult to change, as any change involves manual intervention with the physical components of the network.
A good example of this is the router: in a WAN, a router holds the data being transferred over the network and controls where it should go. If you need to add or change a rule for transferring data, you have to program this rule into the router itself.
This is not only expensive, it requires skills and it is time-consuming – especially for businesses with multiple branches.
Along comes the cloud which is changing the way businesses are operating. Soft phones replace deskphones, more applications, like Office 365, are running in hyperscaled environments like Microsoft Azure and AWS.
The focus shifts from the network itself to the performance of the applications that run on top of it – and the people who use them.
Applications that were hosted on-premises are moving to hosted cloud providers like AWS and Azure. The virtual private network (VPN) that was previously used to connect to those applications can now be replaced by a more cost effective Internet connection, saving businesses a bundle on traditional point-to-point connection.
The SD in SD-WAN
Think of SD-WAN as a wrapper or an intelligent overlay that integrates easily with an existing WAN. It helps tie your corporate network together in a smarter way – and makes it easier to manage.
Now, instead of programming, or ‘hard-coding’ physical infrastructure like a router, you can log in to central online portal for your network, and make changes on a user interface (the ‘front end’).
This doesn’t require you to have specialist resources in every location connected to your corporate network. It’s therefore faster and way more cost-effective.
From failover to resilience
With a traditional WAN: you need two links between points so that if one goes down, you can bring the other one up and operations aren’t affected.
This gives you have an active–passive configuration: one link is active and the other is passive. If the active link goes down, the passive link takes over – we call this failover. You’re paying for two links, but only using one at a time.
With SD-WAN, all links on the network are active, which means you have resilience (and you’re not paying for links that are not using) and you also have the option of utilising more cost effective connectivity mediums like broadband internet to augment or replace the expensive point-to-point links of a traditional WAN.
An intelligent, application-focused network
SD-WAN focuses on the applications that run on your network and makes changes at the application layer – not at an infrastructure level. The software for manging the network has intelligence built in, so it can prioritise traffic for streaming, comms, voice, videos – whatever your users are using. It monitors everything that goes in and out of the network, looks at the health of the links and dynamically switches traffic to the most optimal path.
Think of SD-WAN as the ultimate integrator that turns buzzwords like agility and flexibility into a reality for your network.