The major cloud providers have embraced the push for 5G in the cloud with partnerships and offerings. Learn about the benefits and early limitations of 5G in the cloud.
Telcos and device manufacturers are rolling out 5G technologies that promise faster networking speeds and lower latency. 5G also provides another on-ramp for the cloud providers.
High-speed consumer services have caught the spotlight in early discussions about 5G, but the real promise lies in better connectivity for edge computing. Enterprise architects and developers will be able to build apps that process data closer to the source. And cloud providers will be able to use 5G to fill gaps in their support for these types of applications.
The major cloud providers already have technologies and partnerships to support 5G services and edge computing. However, industry experts caution that it's still early, and there is limited interoperability across 5G services.
"Businesses, along with network and cloud providers, have just realized where 5G technologies can fit in their networks," said George Burns, senior consultant for cloud operations at SPR, an IT consultancy.
Let's explore 5G's impact on cloud computing, which should result in more responsive applications, increased network capacities, and faster media and content delivery.
The edge requires modern app design
The marriage of 5G and cloud computing is in many ways a continuation of what cloud providers already do with content distribution networks used for video streaming. These network services deliver content based on geographic location and rely on distributed servers to cache content closer to end users. However, 5G will support a whole new set of low latency applications that require distributed edge networking to scale.
5G supports a blend of compute, storage and networking at the edge called multi-access edge computing (MEC). Cloud providers, software vendors, enterprises and others will use MECs to run applications directly within the edge of a telco network.
This will require app developers to design applications that can be segmented and deployed across different locations. "Monolithic architectures will simply be too heavy to land out at the far edge practically," said Matt Baker, senior vice president of Dell EMC strategy and planning.
Instead, application components that require real-time analysis will deploy at the far edge. The semi-real-time portions will be deployed in regional or urban facilities, and the non-real-time elements will be deployed at centralized core data centers, including those run by cloud providers.
Emerging microservices and container architectures could simplify app development for edge servers connected to the cloud via 5G. For example, companies such as Liberty Mutual Insurance are evaluating how Kubernetes could extend the same declarative and API-driven development model to edge-based devices.
Cloud providers' response to 5G
To capitalize on the emergence of 5G, cloud providers have expanded their hybrid and edge offerings, partnered with telcos, and built or acquired 5G-specific services. AWS Outposts, Google Anthos and Microsoft Azure Stack are hybrid cloud appliances and services that naturally lend themselves to 5G MEC use cases, said Jim Poole, vice president of business development at Equinix, a leading colocation provider.
Azure Stack and AWS Outposts have limited form factors, which narrows their edge functionality. Google Anthos, on the other hand, is delivered as a software platform that can be deployed on virtually any customer hardware.
Anthos feels more developer-centric, SPR's Burns said. In theory, companies and developers can use Anthos to customize their edge deployments from the ground up. Azure Arc, a newer management service from Microsoft, takes a similar approach.
We haven't seen a major concerted push for multi-vendor 5G test-bed environments to help bring together the hundreds of vendors it will take to make 5G services successful.
Peter DoughertyCEO of MantisNet
Partnerships put cloud providers' infrastructure into telco facilities that could support 5G. AWS partnered with Verizon, Vodafone, KDDI and SK Telecom. It utilized these relationships in the rollout of Wavelength, an infrastructure offering that embeds compute and storage at the edge of the 5G network in major metro areas.
Microsoft made an alliance with AT&T, and it also has plans for smaller, urban infrastructure footprints with its Azure Edge Zones. In addition, Microsoft acquired numerous telco software providers. This includes Metaswitch Networks, which provides cloud native communication software that supports 5G, and Affirmed Networks, which provides virtualized network services that help telcos simplify their networks and launch various 5G services.
Google partnered with AT&T, Verizon, Telefonica and Orange. It recently rolled out plans for a Global Mobile Edge Cloud and added Anthos for Telecom to provide capabilities similar to AWS Wavelength and Azure Edge Zones.
Still, none of these partnerships have produced much utility for users yet.
"While the telco partnerships can seem like a good immediate alternative, only time can tell how [well] these partnerships flourish and if the mobile network operators will be able to scale that capability," Poole said.
Room for improvement
In addition to the changes IT teams will need to make to adapt to 5G, critics are also concerned about limited interoperability. 5G specifications leave some details intentionally loose or open for interpretation, such as the type of cryptographic protection to use.
App developers will need to work closely with cloud service providers to define and develop these offerings and address the fragmented nature of edge-cloud architectures.
"We haven't seen a major concerted push for multi-vendor 5G test-bed environments to help bring together the hundreds of vendors it will take to make 5G services successful," said Peter Dougherty, CEO of MantisNet, a network observability platform.
Most of these initiatives so far have involved closed systems in which vendors and services providers use purportedly open efforts to promote a particular implementation, Dougherty said.
This has the potential to limit innovation and could force technology suppliers and enterprises to choose sides or support multiple concurrent initiatives. This is unworkable for anyone other than the largest players, Dougherty said.
