Greg Pavlik, Senior Vice President and Chief Technology Officer, Oracle Cloud Platform.
getty
When most people think about public cloud infrastructure, they likely still envision clusters of massive buildings stuffed with computing, storage and networking gear. These enormous data center farms, mostly located in a few dozen developed countries, have served as the epicenter of cloud services everywhere.
But in today’s world, businesses and government entities want services delivered faster and from shorter distances. This trend necessitates the proliferation of smaller but still powerful blocks of cloud infrastructure that don’t require vast acreage but can instead fit into a corporate data center or even a secure server room.
The truth is that massive data center farms—a.k.a. “cloud regions”—don’t fit everywhere or suit every need. In 2022, the cloud must be much more distributed to meet these modern demands. It must be available in smaller chunks, in a more diverse range of areas, without losing any of its qualities or strengths.
In this respect, cloud is following in the footsteps of previous generations of computing. In the early days, computers were room-sized behemoths. And, while these mainframes kept plugging away, ensuing years saw the emergence of smaller but powerful minicomputers, followed by desktop PCs, then laptops, tablets, smartphones and even tiny wearables, some of which pack more power than did those early mainframes.
Demand For Data Sovereignty
Demand for distributed cloud comes not just from businesses but from government entities, which need secure cloud infrastructure and may not have a public cloud region within their borders.
Increasingly, some national governments are requiring their data and that of their citizens to remain subject to their laws and regulations. In addition to this data sovereignty mandate, many nations also have data residency restrictions, meaning that their information must be stored and managed in-country.
Such rules pose a dilemma for Gen 1 cloud providers. There are 193 member states of the United Nations, many of which, especially in Africa, South America and Asia, do not host any public cloud regions. What can these countries do to comply with data sovereignty and residency rules?
They can turn to public cloud services that fit into a smaller-than-a-warehouse footprint. For them, there is a need for public cloud services that flow from 10, five or even two racks of computing gear.
Bringing The Speed
Massive cloud data centers strategically located around the world do amazing things, but now, as always, distance matters. The further the cloud region is from the user, the more latency (or delay) in operations. Putting computing power closer to whoever consumes it is critical to eliminating such lags. And that means making cloud computing available in more—and more manageable—form factors than in the past.
But truly distributed cloud services address more than the need for speed. Companies in regulated industries—such as healthcare, financial services and transportation—are required to retain control over some critical operations. That means those companies either have to keep running workloads “as is” on old technology or bring elastic cloud services into their facilities. Putting them “out there” in someone else’s distant infrastructure is not going to cut it.
The Importance Of Parity
One final but extremely important point: Entities wanting to run some workloads in their own facilities and some in a third-party public cloud region must ensure that that the same services are available in both instances. This symmetry enables workloads to “burst” out to public cloud when demands require. If there is no parity between public cloud and private implementations, the mismatch will make such workload balancing impossible.
This is why using a public cloud that makes just a subset of its services available to run on-premises is not ideal and will cause problems down the road.
A decade ago, Gartner distinguished analyst Lydia Leong predicted that public cloud infrastructure would be much more distributed over time to meet these demands. The ideal scenario for many customers is full public cloud functionality available from the site of their choosing, whether a hyperscale cloud region, in their own data center and/or in a server room on premises.
And sure enough, that is what is evolving in the second decade of cloud computing.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
