Do you sometimes get the feeling that purveyors of SD WANs would have you believe that as well as cost savings, instant provision of end point connections and infinite adjustments of both bandwidth and class of service? Claims they can also cure the common cold and arrange for Arsenal to win the Premier League seem just a step away says Ian Hunter.
If only it were so simple; but before I have a go at knocking over some of this hyperbole I’d like to align my current position to being one of ‘this thing could work one day given the right conditions and configurations, but it’s not for everyone and right now and SD-WAN is way over-hyped’.
Can I start with the ‘instant provisioning’ thing because that’s the easiest killer feature upon which to inflict a mortal wound. I’ll make just two points, either of which are individually fatal blows.
Firstly, you have to recognise that SD-WAN is an overlay service that needs an underlay service – the connectivity, upon which to run the service. There are essentially two types of SD-WAN provider; those that provide the overlay service – think in terms of a router or virtual CPE type solutions and there are those that provide a ‘complete solution’ – both the overlay and underlay elements.
In the case of the overlay only supplier, if the SD-WAN controller cannot direct both these elements that means there’s no chance of instant provision or bandwidth flexibility. That feature is in the wind because a third party company controls the underlying connectivity.
My second point here is connectivity specific. How can you provision a node on your network if there is no existing connectivity? That’s quite a common experience I would say in an expanding enterprise location roster.
Well, in our business there are always work-arounds and this problem of no connectivity is no exception. For fixed line connectivity, leased line or internet, the instant provisioning time is directly proportional to the period it takes a supplier to roll up and connect you.
Your experience here will be far greater that mine but I could hazard a guess or three that it would more than a few days. For non-fixed access there is the 4G work-around, 5G would solve the problem but right now… it’s not an option. So, the short term fix is 4G closely followed by the OpenReach van arriving.
Whilst these work arounds could provide a service of sorts, 4G is not omni-present and bandwidth on the internet is both shared and dependent upon distance from the local exchange. I hear you say, ‘What about fibre then?’ to which I would say, you might get lucky – but it’s still shared.
As for the promise of infinitely turning up/down the bandwidth volume control and tweaking the class of service? You can forget about that, and in any case how would you do it anyway – the underlay is in the gift of a third party.
The cost savings argument is perhaps more nebulous I feel. My view is that costs saving are possible but, like most aspects of getting an SD-WAN, highly dependent upon the use case. To keep it simple though I would say that if the client is entirely based in the UK and was willing to risk ‘going hybrid’ – laying off a good percentage of the required connectivity to the internet, cost savings could be made – but not approaching those being universally claimed. Think in terms of a three year ROI cycle right now but remember newer technology will bring down price.
For the use case where a client has a number of overseas operations in an array of countries and continents the cost savings benefits could end up being negative depending upon all the different local charges. Don’t forget to add up all the management costs here either.
SD-WAN is going to be use case driven market for the near term at least and the best positioned suppliers right now will, I feel, be those that who offer that complete overlay and underlay solution.
I’m really hoping for some feedback on this article so please don’t be shy – after all, as my wife and diligent proof reader often reminds me, what do I know?
Come on the Gunners!