Getting a foothold in the 5G applications ecosystem

Application developers have a critical role to play in generating innovative uses of 5G networks, as T-Mobile in the US made clear with its recent launch of 5G Forward to ‘drive developer innovation on 5G’.

T-Mobile put the initiative in place because although today 5G usage generates nearly half the traffic on its network, “the truth is 5G developer innovation has been underwhelming so far”, said Neville Ray, President of Technology, T-Mobile, in the press release for the launch.

Yet there has been no stampede by communications service providers (CSPs) to repeat the attempts made during the early days of 4G to create an application ecosystem. If the industry has learnt anything from 4G, surely, it is that an application’s success depends on it being able to work on any network and the device of a customer’s choosing.

Nonetheless, CSPs can have a key role to play in fostering the 5G application ecosystem and how it impacts their businesses, particularly as they look to support vertical 5G B2B use cases that rely on specific network capabilities.

As Beth Cohen, Cloud Technology Strategist, Verizon, pointed out in a Telecom TV debate entitled ‘Is the edge finally delivering new revenue streams for telcos?’ :“App developers don’t necessarily need to…have deep knowledge about networks, but they certainly need to have the tools that allow them to test their applications and use their applications more effectively and more efficiently over a distributed network.”

However, she also said that telcos have traditionally operated as a closed ecosystem, in which they excel at delivering transport efficiently, while paying little attention to the applications running on top of network infrastructures.

“I think there needs to be more integration between how the network works and how the application works,” said Cohen during the debate. “Network slicing is one way to do it. SD-LAN is another and we’re starting intelligent networks as a third way. All of these are still…not fully mature. But there’s not really full integration between the networks and the developers and the applications and that really needs to happen.”

For this reason, Cohen argues that telcos need to have a stronger voice in the application developer ecosystem than they have had in the past: “The telcos certainly can step up and be smarter about providing tools to the developers. I don’t see the hyperscalers doing that…yet.”

Rakuten Symphony may be trying to address the issue through its acquisition of Robin.io, which aims to “deliver industry-leading solutions for various traditional enterprise opportunities in areas of enterprise applications, data management, cloud and virtualization”.

Ron Westfall, Research Director, Futurum Research, for example, believes Rakuten Symphony’s acquisition of Robin.io puts it in a stronger position to power “the deployment of 5G SA [standalone] networks with their core-based programmability, Open RAN/5G assets and Open API flexibility to allow greater enterprise application development innovation.”

Capitalizing on the edge

Part of what CSPs – along with vendors and hyperscalers – are looking to capitalize on is an evolution in edge computing.

“We’re moving further beyond simply thinking about the edge as being a sort of a low power, small IoT device and thinking about edge components with a full rack and capacity to do full data analytics, and AI training…at the network edge,” says Anthony Behan, Managing Director for Communications, Media and Entertainment at Cloudera.

This change will mean “that you’ve got edge clouds that you need to manage, as well as centralized clouds”, explains Behan. “Technically, from an architecture perspective, you’ve got that challenge, but also commercially, from a business perspective…you’re approaching your enterprise clients and saying to them, ‘you know what, I don’t want you just to connect your devices to my network, I want you to build your applications in my network, and I’m going to give you all of this capacity to help you do that’.”

As manufacturers, for example, combine operations technology with networks and IT to make better use of machine learning and AI, they will generate data that has new transmission and storage requirements.

Dirk Reinert, Lead, 5G-Enabled Campus Edge Solutions, T-Systems, gives the example of computer vision, which is a field of artificial intelligence that enables systems to extract useful information from images and video that manufacturers can use in quality assurance. “You have the camera producing a constant stream of video data that needs to be computed and transmitted. Imagine you’re putting the computer vision on the drone, which is flying through an infrastructure to inspect it and you don’t want any third party to have access to the data.”

Reinert describes this particular usage as “an ecosystem question that is a natural place for hyperscalers to be”, but he also points out that the need for software to deliver a certain quality of service “really is very network specific”.

He adds: “When implementing technology such as artificial intelligence, there are questions around how to organize the transfer of data, and because everything needs to be very flexible and on demand, it all needs to be architected in a way that is really software-defined.”

It’s one of the reasons that Reinert believes “we see that software-defined communication and telco infrastructure are becoming essential parts of the ecosystem as an enabler of digitalization.”

Within this ecosystem, hyperscalers and CSPs can potentially carve out separate roles.

“Hyperscalers have the challenge of creating an offer in their ecosystem, while telcos are providing solutions that will be available across an ecosystem as a kind of standard,” says Reinert.

Hyperscaler ambitions

Hyperscalers, of course, already have well-established application development platforms. And in the case of Amazon Web Services (AWS), the fostering of application development and delivery could link to its 5G private network ambitions.

AWS Private 5G, for example, sets out “to make 5G a much more accessible technology – something that enterprises, educational institutions, factories or warehouses can easily deploy for their private campuses”, explains Bala Thekkedath, Head of Product Marketing, AWS Wavelength, AWS Private 5G and 5G Edge Computing Solutions. “They have full control over their 5G network. It is a private network just for their applications,” he says.

Developers using AWS Outposts don’t have to refactor their applications depending on whether they run on-premises or in the cloud, says Thekkedath. “It’s the same application, the same developer tools and the same development environment. This segues into private 5G, because within that shop floor we can help them set up a private 5G network as simply as setting up a wired or WiFi network.”

AWS is aiming its Private 5G at enterprises that have thousands of connected devices “and expect that number to grow as they continue their digital transformation” says Thekkedath. “And they can install it themselves if they choose to, just like a WiFi network.”

AWS also believes the 5G application ecosystem is starting to take off. “There is definitely a community of software vendors that are looking at the capabilities [of 5G and edge computing]…to create a new set of applications that they bring to market – directly or through partners,” says Thekkedath. “The type of applications that are being developed are completely different – spanning multiple industry verticals and some even disruptive and innovative…5G, cloud computing and edge computing are triggering use cases that were not possible before.”

One example he gives is of UK autonomous vehicle company, Aurrigo, which is monitoring its autonomous vehicles in real time using AWS and Vodafone multi-access edge computing services, which the hyperscaler and CSP are delivering through a partnership.

Meanwhile traditional telco vendors are building on their enterprise network expertise to offer mobile private networks. In the case of Nokia, it is also aiming to make it cheaper and easier to deploy, integrate and manage high-performance computing applications and mission-critical networking. Nokia MX Industrial Edge is designed to give enterprises an “on-premise cloud architecture that unifies edge requirements in an easy-to-use, deploy everywhere, as-a-service package”.

Different doesn’t mean simple

However, as Ian Hood, Chief Technologist, Global Service Provider Business at Red Hat, points out, integration remains a big challenge. “A lot of these technologies at the edge, once you get beyond the bigger sites, don’t really work well on standard computer technology yet.” And even though new applications will be cloud-native, Hood estimates that “80-plus percent of the applications on the planet are still legacy”.

He adds: “The thing that [enterprises] want is to access data and…extract value out of [it]. But that data can be extracted without modernizing the application.”

A large part of any digital transformation involves automating business workflows and processes tied to legacy systems and devices. “It’s automation across all aspects of your lines of business, whether it’s on the IT side, the network side, the operational technology [OT] side, the services side, the back office…all those things need to be automated and…tied together to make this stuff work,” explains Hood.

For this reason, when it comes to answering the question of “how do I marry up that legacy technology with edge compute technology and handle automating the deployment of [it] all?” the answer is still likely to be a systems integrator, believes Hood. “It isn’t necessarily going to get radically simpler even if it changes,” he adds.

Telcos, meanwhile, will retain advantages in both networking and end-to-end application security and are well placed to offer AI as a service or virtual reality services to enterprises, believes Hood.

“It comes down to the application play and how distributed it is from a network perspective. That drives more complexity, which lends itself better to the telco than it does to the hyperscaler,” says Hood.

But picking through the complexity of supporting new B2B 5G-enabled services and applications may not be a quick task.

“The field of potential applications is dramatically increasing,” says Reinert at T-Systems. “And I would say telcos are engaging in that question a lot more. But because it’s more complex it will take a longer time until applications become visible.”