The world of IT is notably susceptible to overblown hype cycles and over-promoted technologies that, in the end, do not quite live up to expectations. One of the big ideas doing the rounds at the moment is the edge data centre. But is the hype justified in this case? Or has other technology simply superseded the need for edge data centres?
The ‘edge data centre’ phenomenon is driven purely by customers’ belief that that the proximity of a data centre to their own sites will increase the performance of their systems. This is certainly an attractive idea, especially for any kind of mission-critical or real-time traffic.
With the growth of video data, cloud-based applications, gaming, and now Internet of Things (IoT) systems – the performance of which are all significantly degraded through higher latency – edge data centres appear to have a huge advantage. With proximity on their side, users benefit from reduced backhaul, lower latency and reduced network congestion.
Certainly, this has been the thinking in capital markets, where data centre proximity to, or even co-location with, major stock exchanges is considered a key advantage to traders looking for sub-millisecond latency.
So, what is an edge data centre?
But as with all hype moments, the original definitions can get a little fuzzy over time. When it comes to the edge data centre movement, there are operators opening or simply re-purposing existing facilities in tier-two or tier-three markets – the areas that have not been well served by proximal infrastructure in the past. These smaller facilities are being positioned as edge data centres that provide low-latency access for local users.
There are also vendors offering small-scale, modular hardware to provide a complete, self-contained data centre designed to operate remotely with minimal support. These too are being positioned as edge data centres – as are certain services from co-location providers. In addition, software vendors are getting in on the game with solutions such as remote management, operations, orchestration and deployment tools made available on the promise that they enable edge data centres.
However, none of this means that a small data centre in a remote region is actually on the edge. When the customers are mainly local and the services being delivered are from local businesses, then, in reality, it’s just a small local data centre. However, it could be providing edge services if, for example, it hosts appropriate hardware from a major content delivery network (CDN) provider. In other words, the edge data centre is not defined by size, speed, service or customer base – the only criteria that count is proximity.
What customers want
But judging by the conversations we have with our customers it seems that edge data centres are more discussed than deployed. Certainly, our customers are more concerned with the IoT than hi-definition video, and in the past few years, the advantages of being able to do more with the data they now have available have driven a growing interest in cloud and networking infrastructure.
But that demand has manifested itself in demand for the centralised cloud-technology model, rather than edge data centres, not least because the cloud has always been about optimising and driving better data practices and making services more efficient.
That the edge data centre concept hasn’t taken off among these customers is largely due to enhancements and improvements to software-defined networking (SDN) that enable the data centre to be extended to a customer endpoint.
In an SDN, the control of data traffic is managed by programmable software and is no longer dependent on forwarding hardware such as routers and switches in the network’s nodes. Administrators can use a central control to regulate the transfer of data and to deliver services to wherever they are needed in the network, regardless of the specific devices, server or other hardware components.
It allows organisations to centrally design, assign and manage application-aware policies and to secure and control all network traffic across all sites. When SDN-clouds are in place, the same concept of software definition is extended all the way from a user’s device to the data centre and whatever private, public, or hybrid cloud arrangements the organisation has in place.
[easy-tweet tweet=”It is important that the network and the cloud infrastructure is aware of the data it is carrying” hashtags=”Cloud,Data”]
Software defined networks – software defined clouds
The fact is that real-time services don’t necessarily require an edge data centre in immediate proximity. For example, a retailer that deploys a software-defined device equipped with IoT sensors and connected via software-defined or a data-aware network can optimise traffic flows coming in and out of that device.
For our customers, the most important thing is that the network and the cloud infrastructure is aware of the data it is carrying and can manage and control all relevant work streams appropriately.
Because SDN technology automates traffic flow and constantly looks for the optimal path through to the back-end, the traditional bottlenecks and barriers to low latency disappear. Combining SDN with the cloud in an end-to-end solution means customers can find new, faster ways of delivering services via devices connected to the IoT – which in turn work much faster and more efficiently than edge data centres.
Underneath all the hype about edge data centres, what we are really talking about is how cloud services are provided. Most cloud service providers are exploring ways to deliver multiple services, across multiple geographies – and across multiple private, public or hybrid clouds. What they need are technology solutions that enable the delivery of applications and services to the locations their customers want. From there, it is about the automation of the delivery of application management and services to those locations.
Edge data centres have the undoubted advantage of being intuitively easy to understand: tell someone that a facility is closer and of course they will believe that it delivers quicker services. Software-defined networking, in contrast, is much more difficult to come to terms with – especially for non-technical business managers with tightened budgets and a Board to report to.
But that’s not a good enough reason to ignore an elegant solution that is much better suited to the world where the IoT, smart connectivity, rapid analytics and real-time information flows are becoming the norm. Quite simply, the combination of SDN and the cloud enable customers to find new, faster ways of delivering services than edge data centres would allow– and so far, our customers agree.
Jonathan Bridges leads the Cloud business at Exponential-e. He has experience in all spectrums of the IT lifecycle including engineering, design and development, strategy, consulting, service delivery, transformation and sales.
In a career spanning the last 17 years, he has worked for major, global system integrators such as Atos, Fujitsu and HP.