Amazon, Google and Microsoft sit on top of the hyperscale ecosystem, and each of them is staking their claim in 5G
Managing networking and Internet service infrastructure is an essential function of the cloud ecosystem. Getting that infrastructure to scale elastically to meet the often opposing demands of consumers and enterprise requires hyperscale services.
Hyperscale’s apex predators
Hyperscale computing describes data center servers and operations optimized to rapidly deliver cloud computing services. The market for hyperscale services has exploded with the growth of the cloud economy. The mid-tier market is rich with competition from regional players, smaller cloud providers and Software as a Service (SaaS) providers. But the businesses at the top of the hyperscale market operate millions of servers in worldwide data centers with vast amounts of interconnection.
AWS, Microsoft Azure, Google Cloud, Alibaba and IBM all provide public cloud hyperscale capabilities and comprise 80% of the public cloud market, according to Synergy Research. Other companies like Facebook and Apple operate data center operations at comparable scale, but specifically for their own operational needs.
Hyperscale and 5G: collaborate or compete
The telecom industry is transitioning to 5G Core Network technology defined by a Service Based Architecture (SBA). SBA is the same service-oriented architecture design principle that enables cloud services to work and to scale to meet public demand. Hyperscale operations are absolutely necessary for Communication Service Providers (CSPs). But how the current hyperscale market fits into the carrier’s vision of the future is still to be determined. Hyperscale services are a necessity to carriers and a threat to future business.
CSPs are understandably wary of hyperscalers. When it comes to edge computing, for example, most carriers perceive hyperscalers with an unfair advantage in the 5G business, according to a report from IBM.
“CSPs shoulder most of the burden of investing in infrastructure, but most money is being made by using that infrastructure — not building it,” said IBM.
Hyperscalers have been able to thrive in an environment absent of the restrictive regulatory environment imposed on carriers. That burden is also where some in the hyperscale industry see the biggest built-in advantage for carriers.
“They’ve got the fiber out there…they’re already in everybody’s home. The hyperscale web guys don’t have that. They might try to build it like Google. But it takes time to build that kind of infrastructure everywhere throughout the globe.” said Jyeh Gan, director of Dell EMC’s Extreme Scale Infrastructure business, in an interview with RCR Wireless.
Hyperscale at the 5G edge
Hyperscalers have been quick to offer Multi-access Edge Computing (MEC) solutions that push compute functions from data centers to the edge of the network, closer to the user. MEC is a pre-requisite for the high-bandwidth, low-latency functionality that advanced 5G Standalone (SA) networks offer. There’s room for collaboration and competition as everyone lays the groundwork for long-term 5G business growth and expansion.
Verizon and AWS are expanding their MEC service offered in major U.S. cities. The service moves AWS compute and storage services to the edge of Verizon’s 5G Ultra Wideband network for customers developing ultra-low latency applications supporting next-generation use cases: self-driving cars, autonomous industrial equipment and more.
Google’s Global Mobile Edge Cloud (GMEC) vision sees the company helping carriers to monetize 5G, particularly as a business services platform beyond the consumer market; using AI, ML and analytics to develop “data-driven experiences” for customers; and improving operational efficiencies by using the cloud for their OSS, BSS and other systems.
Hyperscalers have a role to play in network disaggregation as well. Nokia is embracing partnerships with all three dominant hyperscale players to accelerate enterprise 5G adoption. Nokia has partnered with Google Cloud, AWS and Microsoft on Open RAN and Cloud RAN edge computing efforts.