At the Red Hat Summit in Boston, I met my old writing buddy from 1993, David Linthicum to discuss where the cloud is in 2023. Now Deloitte Consulting's Chief Cloud Strategy Officer, Linthicum sees the cloud gaining ever greater importance and complexity.
For starters, Linthicum sees more specific, industry-targeted cloud services emerging. While these services may operate on major cloud provider infrastructure, they often function as separate cloud entities themselves.
This growing network of cloud providers harkens back to the era around 2009-2010 when the number of viable cloud providers shrunk from around 40 to just three hyper-cloud providers – Amazon Web Services (AWS), Microsoft Azure and Google Cloud Platform (GCP). Linthicum observed, "This is kind of funny, but now we're kind of building things up again to deal with specific issues such as AI engine services."
Simultaneously, smaller players "that may have missed the initial cloud boom, like Oracle, IBM and Red Hat, are stepping up to offer abstractions and automation services that can go across multiple clouds." For example, Red Hat's new Red Hat Service Interconnect enables users to deploy and manage application connectivity and security across in-house platforms, clusters and clouds.
Business use of multiple clouds, whether you call them superclouds or metaclouds, are, Linthicum stated, becoming ever more important. Whatever name you use – I like Bigly Cloud myself – it's how businesses use clouds today.
But Linthicum warned this creates “complexity problems." For example, "instead of running one security system across cloud providers, they're running seven Still solutions that can now mitigate that complexity, guys like Dynatrace, BigPanda and cross-cloud security players.
He continued, "We need to improve how the various cloud providers and the existing on-premise legacy systems work and play well together in ways that bring more value back to the business. The cloud providers need to stop building silos and walled gardens. That's not the reality moving forward."
Navigating a multi-cloud future
So how do we get to this new multi-cloud reality? Linthicum thinks the answer is Kubernetes.
This, he said, "will become the platform of the multi-cloud. Eventually, we will get to where we can do development and ubiquitously deploy application components, aka containers, which simply use the resources they need regardless of the underlying cloud." In this world, "cloud providers become a commoditized service. The federated containers and the container orchestration infrastructure run on whatever."
In this future, Linthicum believes we'll "be able to deploy applications to a federated Kubernetes layer control plane, and it will run the containers on whatever resources are cheapest and best for the job. This will commoditize the cloud providers, but I think that's a good commodity." Needless to say, the hypercloud providers may disagree.
Linthicum is also concerned with how we'll use artificial intelligence (AI) to address the increasing complexity of cloud services. While AI has the potential to automate and abstract various systems, too many companies are "AI washing," where companies list AI in their features but don't really exploit its capabilities.
At the same time, one should keep an open mind about AI and understand what its problems are and take an objective look at what AI can do and how it should be applied. Just don't assume that an AI program that promises you the world or can actually solve all your cloud problems anytime soon.
SaaS and edge
We also discussed other cloud aspects. For example, Microsoft has long been shifting from PC-centric Windows and Office products to Software-as-a-Service (SaaS) approaches such as the Azure-based Windows 365 Cloud PC and Microsoft 365, formerly Office 365. While Linthicum couldn't talk directly about Microsoft's efforts in this area in detail because Deloitte refrains from commenting on specific vendors, he did comment that generally "Things have been moving to kind of a SaaS-based delivery, on-demand delivery for some time. Ultimately, anybody who has things that were normally localized on a desktop is going to move to a consumption-based cloud system."
This also means that edge computing, though more of a marketing term, will continue to become more important.
"To become efficient with your cloud architecture, you must quit sending everything 5,000 miles away when you can collect data on-premise or on the edge and do the backend system processing as needed close by,” he said. “To me, the edge is an architectural evolution that has pragmatic advantages that people should always think about."
To sum up, we're entering a period of great and fast cloud evolution, powered by AI, which will also bring enormous funds into the cloud space. This transformation will lead to commoditized, Kubernetes-based superclouds, where both our server and desktop applications will work.