Telcos often view hyperscale cloud providers such as Amazon Web Services, Google and Microsoft as competition, complaining that the digital natives have a distinct advantage because they aren’t hamstrung by regulation. But these companies are also potential partners for communications service providers (CSPs), and they could help them crack the enterprise 5G market.
That’s the message Microsoft’s Eric Troup, Chief Technology Officer, Worldwide Communications and Media Industries, delivered during a digital leadership session at TM Forum’s Digital Transformation World last month.
“Leveraging 5G to deliver the next generation of enhanced user scenarios is a challenge we [Microsoft] cannot solve alone and the service provider can’t solve alone,” he says. “It really is a partnership discussion.”
Troup adds that this is not necessarily a technical discussion – it is often a business-level conversation between CEOs: “Oftentimes these deals are happening at a higher level, and that helps drive consistency across both enterprises. It helps us prioritize what we’re trying to do.”
Moving power to the edge
Microsoft is addressing consumer and enterprise 5G use cases by enabling cloud capabilities, including artificial intelligence (AI) and machine learning (ML), internet of things (IoT), video processing, facial recognition, game streaming and mixed reality, closer to the end user in order to optimize the experience. Troup talks about this in terms of the “Intelligent Cloud” and the “Intelligent Edge” connected through 5G.
Reusable building blocks
Microsoft runs one of the largest IP (internet protocol) packet networks in the world, operating four discrete platforms – Microsoft Gaming, Microsoft Dynamics 365 & Power Platform, Microsoft 365 and Microsoft Azure. These “Platform Clouds” provide reusable, building-block services consistent with TM Forum’s Digital Services Reference Architecture (DSRA), which is now part of the Open Digital Architecture.
“Azure provides a large set of reusable building blocks that enterprises can use to build and operate their business platforms,” Troup explains. “These can then operate over the Intelligent Cloud and the Intelligent Edge.”
In the Intelligent Cloud, for example, the Microsoft Azure environment consists of 54 global regions, each with two or more data centers that are up to a square kilometer in size. “The perimeter from one side of the region to the other is about two milliseconds, and the entire environment is a software-defined everything,” Troup says.
The Intelligent Edge encompasses millions of edge devices and hundreds of edge data centers. It is becoming increasingly complex as the edge expands. This is where telcos can play a crucial role delivering 5G connectivity and value-added services such as data analytics, security and quality-of-service guarantees.
“We are not going to compete by building an edge network with 5G – we have no plans whatsoever to do the last mile,” Troup emphasizes. “We’re not building the 5G edge. We are using it – we need it to work – but we’re not building it.”
He adds: “As we look at various user scenarios, we’re going to need to connect our two software-defined worlds together using different data paths. If I’m moving email, it might be Data Path 1 because it’s traditional IT data – there’s no low-latency requirement associated with it. However, many of the newer scenarios require less than 10 milliseconds latency from Microsoft Azure. To meet these requirements, Azure may need to peer with the telco network at the edge, possibly in a central office or beyond. For instance, high volumes of facial recognition or managing autonomous cars may require Data Path 4.”
The Microsoft Azure resource-management environment is dynamic and AI-driven. Load-balanced workloads are moved around seamlessly from one location to another based on real-time data analytics. The idea is to increasingly get them “as close to where the action is” as possible to reduce latency and improve customer experience, according to Troup. Combining this with 5G connectivity requires partnerships.
He explains that there are two types of edge computing that are part of the Azure platform infrastructure: network edge computing (NEC – a term AT&T coined and used in a press announcement at MWC 2019 about its work with Microsoft to develop 5G enterprise use cases) and multi-access edge computing (MEC). As described in the DSRA, various “tenant” applications request use of this platform infrastructure using Open APIs to enable business use cases. For example, an IoT workload will have different requirements for latency and availability than a workload for IT, gaming, media or network virtualization.
“You grab capabilities from one or two or three of those categories, and you build your business platform,” Troup explains. “It doesn’t necessarily get deployed in one spot – that’s the dynamic nature of the actualization platform. You’ve got your hyperscale, the telco or enterprise that have data centers or central offices, and you also have the edge device.”
"The point is to build an environment where optimization is possible – and to let partnerships be driven by use cases."
Troup gives examples of an enterprise that wants to offer a smart retail experience or a company that wants to use Microsoft HoloLens to perform remote maintenance of equipment. These scenarios could require IoT capability, video processing and data analytics at the edge to get the latency down to within a millisecond, and they require applying AI and analytics securely to users’ data with guarantees of privacy (Troup stresses that Microsoft does not monetize its customers’ data in any way).
“We’re creating a joint value proposition to solve a specific enterprise customer problem,” he explains. “So, rather than saying, ‘Okay, I’m going to build a 5G slice and then I’m going to figure out how to monetize it, what we’re doing is starting with an enterprise customer that has a real business case.”
Leveraging private 5G
In May Microsoft and SK Telecom signed a memorandum of understanding to combine their technological capabilities in areas such as 5G, AI and cloud to jointly promote IoT business including smart factory; AI technologies and services; media and entertainment services; and new ways of working for ICT companies under the SK Group umbrella.
“They’re moving their 5G and all of their other operations to the cloud environment,” Troup says, which will help both companies learn how to create and support joint value propositions.
Ultimately, Troup envisions using the experience gained operating private 5G networks (Microsoft is building its own private 5G network as part of its campus modernization project in Redmond, Washington) to define requirements for the orchestration, operation and management of services that leverage the Intelligent Cloud and Intelligent Edge connected with 5G.
“We’re using the limited scope of private 5G to figure it out because there aren’t any real, written-down requirements,” he explains. “So, we’re building it – run-break-fix, run-break-fix – making sure we get edge computing (MEC/NEC) right.”