After years of talk and hype 5G is starting to emerge as something of a reality. In February, the European Horizon 2020 Project launched its testbed for trialing futuristic media services in Bristol and Barcelona, while the latest 5G spectrum sell-off saw UK operators fork out nearly £1.4bn in April. US operator AT&T announced it will launch a 5G network later this year, while a report by Viavi revealed that there are now 72 operators across the globe currently testing the technology.
There are no shortage of trials and according to a recent CCS Insight report, it is the US and Asia that will dominate early adoption. This won’t of course stop cities across the globe claiming 5G firsts (the University of Bristol demoed the world’s first end-to-end 5G network at MWC in March) but what it means is that the technology should mature rapidly.
We’ve already had our first commercial taste at the recent Winter Olympics in Pyeongchang, South Korea, where Intel, Samsung and KT collaborated to provide high resolution streaming and VR experiences, thanks to 5G-connected cameras. So, we know it can work in the real world at least on a small, controlled scale but what we really need to understand is how this will impact computing on a wider scale.
Japanese giant Toshiba aims to provide a platform for making integration between the edge and the network easy. Find out more: Toshiba has its eye on the wearables and edge computing space
There is a coming together of new technology. 5G is maturing at a time when demand for connected devices and sensors is increasing and with it pressure on networks to manage increased amounts of data. This is where edge computing steps in, to increase processing and analytics capabilities nearer to the source of the data. It’s certainly gathering momentum, with Intel announcing a chip designed for computing at the edge back in February. What it means is that to cope with increased complexities and the demand for real-time monitoring of networks and datacenters, organizations will have to look for more automation.
“Smart companies are pushing their compute and analytic capabilities across their IT landscapes to the edge and finding that it serves growing needs around real-time analytics and supports taking action at the speed of your business,” says Shawn Rogers, senior director analytic strategy at Tibco. “Moving analytics to the data instead of the traditional approach of moving data across the network to the datacenter saves time, reduces complexity and allows companies to automate actions and insights, mash up valuable data and drive ROI.”
Rogers actually goes on to say that for any business to stay competitive, maintaining a solely centralized strategy is “antiquated and likely to reduce your ability to digitally transform your company.” He has a point. It’s a view shared by Carl Grivner, CEO of Colt who says that as devices multiply, the computation needs at the heart of the Internet of Things will increasingly shift to the edge.
CIOs are now used to cloud computing, but here comes edge computing. Should they worry? Find out: Edge computing 101: A CIO demystification guide
“Maintaining low latency is one of the main reasons companies are moving large amounts of data from IoT devices closer to cloud processing and analytics at the edge,” says Grivner. “Placing interconnection at the edge will also save on network costs, as companies filter out volumes of useless IoT data near the source to gain faster access to valuable insights needed by IoT-enabled innovations like smart hospitals.”
It is this capability that is accelerating the development of IoT ecosystems. As Chris Wilder, Moor Insights & Strategy senior analyst wrote in a recent post, a number of large ecosystems are already developing from major brands such as Amazon Web Services, Cisco, Dell and IBM. It’s becoming clear that 5G and edge computing will be pivotal to the further development of these ecosystems, but we are still a long way from being out of the woods in terms of challenges.
Challenges
For one of course it’s still early days on 5G and edge computing. While 5G and IoT are quickly expanding the scope of computing beyond the datacenter, how businesses manage that process is still evolving. While the need for automation is an obvious answer, how that automation is implemented still poses challenges both in terms of cost and capability. There needs to be interoperability between ecosystems to ensure that businesses are not forced into any proprietary cul-de-sac but ensuring a smooth transition to centralized network management takes some planning.
“It’s commonplace for companies to manage several types of compute engines (stream, algorithmic, edge) but applying them at the right point in the process is the innovation tipping point,” says Rogers at Tibco. “Distributing analytics, AI and ML to the edge is the new normal. Controlling, optimizing and leveraging these environments will place new roadblocks in the way of edge adoption.”
Enterprise firms jostle for share of burgeoning IoT space as computing moves to the edge… and Dell aims to lead with new division and strategy. So What does $1 billion buy you as IoT moves computing to the edge?
Rogers adds that centralizing the management approach, so it’s easier to maintain control is critical. As the landscape expands so do challenges with security and overall agility.
“Without the proper plan and solutions in place growing to the edge could become an inhibitor to innovation instead of a catalyst,” he says.
It’s a warning shot.
John English, senior manager, service provider solutions at NETSCOUT says that enterprises need to select the right carrier services providers that have the agility to rollout and manage these new networks and services. As always, buying in expertise is crucial but which experts?
“The right CSP partners will have the visibility and ability to achieve an accurate view of what’s happening across the 5G gateways being effectively a geographically distributed remote datacenter, which will require service and security assurance across the network,” says English. “Having visibility and intelligence into these new cloud-based edge nodes and service delivery infrastructures is absolutely critical to the successful rollout and ongoing service delivery.”
The problem at the moment is that we are not really sure what patterns will emerge once the IoT network is up and running, so any anomalies or vulnerabilities are going to be harder to identify.
“Network edge computing demands enhanced service and security assurance, to provide visibility into virtual resources,” adds English. “This will facilitate not only proactive and reactive monitoring and trouble-shooting, but also the move to predictive and prescriptive paradigms, with real-time smart data feeds for automation and orchestration to achieve self-optimizing networks.”
He adds that the analysis of the monitored data provides end-to-end visibility at an IoT service level, as well as across the service delivery network. This of course, will help enterprises to overcome the challenge presented by edge computing, by mitigating risks, accelerating business agility and promoting operational excellence.
So, while 5G promises ultra-low latency and increased bandwidth, it can facilitate growth and manageability of a wide range of IoT devices and services including connected/autonomous driving vehicles, smart cities, emergency/security applications, and power sensors. It demands a datacenter at the point of data collection but this is just the start of an evolving future of multiple network scenarios to cope with the varying urgency of data analytics and corresponding computing action.
“As datacenter operators look to benefit further from economies of scale, and many of the older datacenters are at capacity, we’re seeing the adoption of hyperscale datacenters, with considerable power, cooling and connectivity requirements,” comments Grivner at Colt. “When data is created it is typically backhauled to an aggregation point – such as a datacenter – where it can be stored and manipulated and perhaps accessed. But it’s expensive to transport data long distances and in some cases, there is value in being able to manipulate that data in near real time. As a result, it doesn’t make sense to move it to some distant aggregation point. So, edge storage and compute requirements may well see datacenters move back out to the edge.”
As the IoT continues to penetrate many aspects of everyday life, including critical applications such as healthcare and autonomous vehicles, so this scenario becomes increasingly essential. Clearly 5G is a catalyst but it’s only part of the puzzle. There will be different horses for different courses, as English at NETSCOUT suggests:
“IoT devices will likely segment into different classes of service – from critical to non-critical – and carrier service providers will need to offer service level agreements for those critical services, to guarantee network and end-user satisfaction and, more importantly, security and safety.”
Also read:
Edge computing 101: A CIO demystification guide
IoT and 5G are driving computing to the edge
What edge computing means for business IoT
IoT set to push computing to the edge in 2018
What does $1 billion buy you as IoT moves computing to the edge?
AWS wants to lower the bar of entry for all companies wanting to do IoT
GE adds edge analytics, AI capabilities to Predix industrial IoT suite
Toshiba's in chaos, but not quitting PCs -- yet
Epson a steady presence in Augmented Reality
Are smartglasses due a “phase 2”?
Why France may lead in the Internet of Things