One of the issues I have with much of the PR hoopla around SD-WAN is its tendency to major on the cost savings, with the implication that organisations could finally unshackle themselves for good from the chains of expensive MPLS connectivity. So is it as they suggest, that MPLS is on borrowed time, given the efficacy of the SD-WAN alternative?
In the short-to-medium term I don’t believe MPLS is going anywhere: it will continue to be trusted as a premium, performance-based, security-centric transport, and will remain a fundamental component of most enterprise networks. But organisations will certainly use it more judiciously, and that may bring some savings along the line – but that’s not really the point of SD-WAN, and to just make it all about ‘cheaper connectivity’ does it a real disservice.
What we should be focusing on is its ability to modernise networking. Leading vendor Riverbed talks about the network needing to be “a strategic asset that emulates the Cloud — elastic, dynamic, and software-defined. But the enterprise network of today is out-of-date, brittle, and device-centric…it’s an incredibly complex, insecure, and unpredictable environment.”
SD-WAN has a pivotal role to play in addressing that challenge: re-architecting networks for the Cloud-connected era, and driving significant customer benefit as it does so. At Redcentric we are already seeing two strong use cases emerge – neither of them driven by a rush to abandon MPLS, rather both of them making strong operational and commercial cases.
First, it gives us the option to move clients to a co-management model. Co-management is not us giving up our ‘managed service’ work, it’s sharing tasks and responsibilities in a way that better aligns with the needs of a client; it’s also taking advantage of the very obvious efficiencies inherent in SD-WAN, efficiencies that accord well with growing demands for agility and responsiveness.
Let’s take the case of a 100-site UK retailer with a managed network service. They want to deploy a new network device to a location, meaning they need to request an engineer to visit the site, install the device, configure and test the connection before anything else. That is then added to the other 99 devices that now need administering: if there’s a policy change, the retail client needs to make another request for those 100 devices to be updated via individual CLIs. Those changes will always be of a lower priority than service-affecting tickets, and the task itself is time-consuming and laborious. It’s the very opposite of agile, efficient, responsive; it’s far from optimum but up to now it’s been the accepted standard for managed networks.
What’s exciting about SD-WAN is that it allows us to get much closer to that optimum delivery. We can hand back control of basic network administration without any loss of overall oversight or erosion of our supporting role; we just stop being the bottleneck for adds and changes, and the client starts being the orchestrator of WAN performance and policy.
So for our retailer, there are no more requests for an engineer. Instead, they use SD-WAN’s zero-touch deployment capability: following some initial design work for the network node via a central console, an SD-WAN appliance is shipped to the store and plugged in. Following discovery and identification, the central SD-WAN controller automatically goes through the steps to bring the node online, configuring it, applying policies etc. No further intervention needed.
And that same speed and frictionless implementation is mirrored in on-going policy management. Forget a change request ticket; a policy change, for example, an uprading of security requirements or prioritising of certain application data, this change can now be quickly instigated via the policy engine of an SD-WAN controller. That in turn will ensure the new rule is downloaded to all SD-WAN devices at the click of a button. What routinely could take days and weeks has now been reduced to a matter of seconds. And for our retailer and many hundreds like them, who are expected to be agile by default in today’s world, that ability hasn’t come a moment too soon.
The second use case is perhaps a less obvious one but it’s a brilliant solution to a previously intractable solution. Organisations are merging and acquiring at an unprecedented rate; from a technical operations perspective, that often leaves the organisation – and its managed service provider – with the headache of legacy and redundant networks. How that plays itself out is often highly unsatisfactory for both parties – early termination penalties on one side, compensatory commercial arrangements on the other, a network rework that can be an expensive rush at risk of compromise in that urgent push to get to a unified, streamlined WAN with no loose ends.
What we’re finding with SD-WAN is that it can take off that pressure and allow everyone to breathe and to transition in a more measured way. In place of a mish-mash of disjointed technologies, connectivity options and existing network equipment from multiple vendors, you can place an SD-WAN device in every location and architect through software a uniform network of networks.
It’s a very neat solution to an invariably messy problem. It buys time, it buys control, it buys peace of mind. And yes, it may also save you money but don’t let that blind you to some of SD-WAN’s even bigger benefits.
For more information, contact the team on 0800 983 2522 or email email@example.com.