Rewind 30 years, the 1980s - an era of terrible fashion, big hair but the start of organisations embracing technology driven by the desire for connectivity to the distributed estate, linking offices, branches, and retail outlets adopting point-of-sale systems. It was a time of low bandwidth requiring high security to protect the passing of sensitive information to and from the data centre; initially deployed over the network by X.25, then ISDN, with things really starting to gather pace with the evolution of MPLS.
Over time, building wide area networks based on point-to-point links, these networks became complex, messy, and quite expensive in relation to the value that was ‘added’ and, over long distances, prohibitively expensive.
A key advantage however of MPLS over previous technologies was that it enabled you to have a private network without having to manage all of the complexities of the individual links. A service provider could provide a link from a central point. You didn't need to worry about having a link from building A to building B. It was building A to the central point, building B to the central point. And so MPLS became, and has been the mainstay for private networks for over 15 years; providing connectivity that is exclusive, secure and distinct from any other sort of connectivity.
The next ‘real development’ in the wide area network (WAN).
Fast forward to today; the way that organisations engage with their staff and with their customers is all driven by technology - we’re all walking around with a device that is equivalent to what most end-point computing pieces were 15 years ago. The level of connectivity and the amount of bandwidth needed is increasing at an unprecedented rate putting the wide area network under a huge amount of pressure.
If you have a private network that is relatively static in the amount of bandwidth that it provides, then uplifting that bandwidth can be quite complicated because there tends to be tiering points about how you access your wide area environment that drives the cost base of what that is. If you're a largely distributed organisation you will historically have had a connection of X amount of megabits per second. Once that starts to come under pressure then you are potentially staring down the barrel of some significant cost uplifts to be able to provide the bandwidth you need.
There are two options at this point, you can either accept those cost uplifts or delay the innovation, which means that you can’t deliver the project because of the prohibitive cost of the network connectivity. And until now that's really been the reality of where organisations are. So to a certain extent, the wide area network has been a handbrake on innovation and what can and can't be done in the distributed environment.
So how do I take the handbrake off innovation?
This is an extremely pertinent question right now as our industry, and certainly the service provider industry would have you believe that you need ‘bigger, fatter pipes’ encouraging customers to increase their wide area network bandwidth typically via something like MPLS.
There are a number of different mediums that you can connect to a service provider's core network. When that starts to reach ethernet or fiber, then in an MPLS world it starts to become very expensive. Since organisations are looking to use technology in different ways, and have ever increasing requirements for video, and voice, the volume of applications that those sort of mediums can be used in, and not forgetting the applications themselves, then actually what we see is organisations hurtling towards those levels of bandwidth requirements.
We are also seeing organisations building applications without properly considering the impact on the network and this can be a nasty shock when they realise what they then need to do in the network, and it usually won’t have been budgeted for. It’s a problem, and that's where an organisation like Gyrocom can help.
There are ways to overcome this efficiently with some of the propositions that we work with. There is a requirement for certain things across your wide area network, and you have to be able to provide headroom to be able to do that. There are clever things that you can do, but there are also some natural laws that you need to figure out. Our experience is that there is a lot of confusion about what those natural laws are, and the bigger, fatter pipe mentality isn’t necessarily always the answer.
So what are the alternatives?
If you're in an environment where your application is latency intolerant, the bigger, fatter pipe mentality is not the issue. With voice and video for example, you can't wait for a round trip between point A and point B – it’s “real time”- if you say something the person on the other end of the line needs to say something back conversationally. So the issue with those sort of applications isn't the amount of bandwidth that you've got (albeit you need enough bandwidth to be able to support conversation or a video stream) the controlling factor is latency. It's the delay between one end of the conversation and the other end of the conversation.
Now when you start to overlay voice, video, email, and other applications, you have quite a complicated mix of what your traffic profile and requirements look like. How do you address that? It’s an involved conversation that requires specialists. The bigger, fatter pipe approach is like using a sledgehammer to crack a nut; you have to be more sophisticated than that. Plus over the last 15 years we have some quite sophisticated techniques that can improve the quality of delivery over smaller links, and perceived inferior links. There’s the perception within the industry that, for example, a DSL connection isn't as good as an MPLS connection. Certainly the cost points are fundamentally different, but it all depends on what you're trying to do with that link. As an analogy; if my objective is to get from A to B, I can achieve that objective in both a Mini and a Bentley. However, if my objective is to get to A to B in style, then the Bentley is the thing, a Mini still gets me there though.
As pressure grows on the WAN, it’s necessary to become granular about what your actual requirements are. In order to do that you need visibility, and I’m afraid that this is where we are very under-serviced by the service providers, because they really don't want to give you the visibility of what's going on at the core of their environment. It’s their domain based on the objectives that they're looking to drive, which is the bigger, fatter pipe, and the bigger revenues from the private network.
So what we've seen over the last 18 to 24 months is the growth from a market area called software-defined wide area network. This is a description of a set of technologies that allow you to create the equivalent of a private network but define it in software where the underlying physical connectivity can be made up of multiple different types of connections. SD-WAN is a set of techniques and technologies that are being brought together in a meaningful way that have actually been around for the last 10 years. They've grown out of things like WAN optimisation, application acceleration, and link bonding, and have been brought together in a meaningful way in order to address the challenges that organisations are experiencing around increased bandwidth requirements.
It’s important to recognise that SD-WAN has grown from literally nothing to being a market sector that now has in excess of 30 technology vendors delivering into it. The function of SD-WAN however, is not a vendor conversation, it's a capability conversation. And significantly it provides an alternative (or ancillary benefit) to MPLS. With SD-WAN I can take my existing provision from my service provider and I can augment that existing provision with bandwidth from any other service provider, or even the same service provider, or point-to-point connectivity, or mobile 3G and 4G. I can create that environment and give myself a superior experience over the top of that environment that I get from any service provider.
So what's essentially happened there is the service provider community and the services that they provide have been commoditised; there is no secret source anymore. What I need is a commodity connection and I can build the sort of environment that is appropriate for my requirements overnight. In addition to that, if I do that I get all of my visibility back. I can see everything that's going on within my environment, which allows me then to get much more efficient in terms of what sort of connectivity I put in place.
And what we see is organisations that grasp this concept have the ability to be able to remove the reliance on the service provider, or a single service provider, and actually build much more appropriate networks, much more cost effective networks with a lot more control and a lot more visibility. And in a world where you don't know what the next application you need to deliver is, that's hugely important.
Discover more about SD-WAN here.