At the dawn of 2020, before anyone had any idea thatwould soon change everything, we ventured out our year-ahead predictions. We took stock of the growing interest in the cloud. Obviously, we're not the only ones who've noticed. We also took note of how the use cases evolved from tactical to strategic and came to the conclusion that in the 2020s, the default choice for deploying IT systems -- entrenched and new -- would flip from on-premises to cloud.
We termed the new era, not as "The Cloud Default," but "The Hybrid Default," because, as enterprises got beyond those obvious born-in-the-cloud use cases to the foundational systems that keep the lights on, that their cloud decisions would have to grow more nuanced. They want the operational simplicity that cloud-native deployment can convey. In some cases, the public cloud would be the right path, but they would need to find a way to make it work on-premises for many cases. And that's where we pick up the tale of IBM.
Turning lemons into lemonade
IBM was admittedly late to the cloud, with the big three public cloud providers having stolen the thunder. But now, IBM has the opportunity to turn lemons into lemonade. There are customers who don't want to get tied into specific public clouds, and as noted, there are customers requiring options that bridge to their data centers. And so IBM has found its beachhead in hybrid. But where is the most promising starting point?
First, let's zoom out. With Arvind Krishna having taken the helm as CEO barely a year ago and the ink from the $34b Red Hat acquisition now drying, IBM is clearly at an inflection point. The big themes around IBM today include AI; Quantum Computing, what becomes of its global services business, and, of course, the cloud. It's the latter that we'll focus on here.
Just stating the term "cloud" does not adequately characterize IBM's cloud strategy. It's not just "cloud," but hybrid and multicloud. Yes, IBM has its own public cloud, and as we'll note below, for part of its portfolio IBM Cloud will play a strategic role.
More to the point, IBM is telling its customers that if you already have begun your cloud journey, it's nowhere near complete and that the heavy-lift is dead ahead. You've already started the journey with DevTest, then new apps such as mobile, edge, and customer engagement. Maybe your teams have also dabbled with some of the new AutoML services that are popping out there.
But what about the core enterprise systems that keep the lights on? These are the systems that are subject to strict regulation; the applications that have too many local interdependencies; or the data that can't move to a public cloud because of some internal policy or legal restriction. This is where cloud transition presents unique challenges, and where enterprises might have to take different approaches to that transition compared to earlier forays with born-in-the-cloud mobile or edge apps.
Then there's the question of cloud lock-in. It's one thing to run DevTest workloads in a cloud because those workloads are transient and, in most cases, are able to run in any cloud where the price and availability of compute instances is right. But running something like an ERP or core banking system in the cloud is going to be a more strategic decision because those workloads cannot be quickly moved. Here, choosing a cloud provider will be like making the Oracle or SAP decision all over again. It's deciding who, and at which level of the stack, will be the strategic technology partner: will it be the cloud, application, or database provider that defines how and where the workload will run?
Cue up IBM and Red Hat. Two reasons stand out as to why IBM paid $34x revenue for this $1 billion company. First, with the aspirations of making it a reverse acquisition, IBM hopes it can become more like Red Hat, and as part of that transition, spun off part of its GTS business last year to help get it there. And secondly, for OpenShift, it enables IBM to become cloud-agnostic. OpenShift is portable Kubernetes (K8s) based software designed to run in any cloud, public, private, or hybrid -- meaning that theoretically, modernizing on-premises systems does not necessarily have to result in a shotgun marriage to any specific public cloud.
There are lots of K8s platforms out there, but they are not all alike. There are differences in how each platform provider implements administrative functions like multi-tenancy, service mesh, role-based access control, encryption, and so on. OpenShift is currently available on AWS and in preview on Azure. Meanwhile, Google Cloud is making a similar play with Anthos, and ditto for Microsoft with Azure Arc, with the difference between them choosing to deploy with K8s or conventional VMs. VMware and others are also offering their own K8s implementations.
Then there's the question of who deploys and manages these K8s environments.
Anthos and Arc put you on your own; your team needs to know how to set up K8s clusters, not to mention knowing the ins and outs of deploying them in different clouds. This path ain't for beginners. At this point, few organizations have the requisite K8s skills or attendant knowledge about creating microservices architectures and setting up service meshes.
This is where IBM/Red Hat differs. OpenShift comes in many shapes and sizes. Customers who know their way around K8s can do it themselves on any hardware, just like with Anthos or Arc. But you could instead take advantage of Managed OpenShift, where you still need to know how to set up containers and microservices, but leave the management of the physical implementation to IBM Red Hat. Or you can go further and get a platform as a service, such as Cloud Pak for Data, where you focus on the analytics and machine learning models rather than K8s, with all the Kubernetes plumbing safely tucked under the hood. And then there's IBM Cloud Satellite that has K8s under the covers, but is in fact a prepackaged extension of the IBM Cloud that runs on-premises, which just went GA a few weeks ago.
Vertical clouds could be IBM's sweet spot
There are lots of providers making sounds about multicloud beyond IBM or at least two of the public cloud suspects, Azure and Google. Numerous software-defined cloud or cloud-like providers from VMware to HPE, Nutanix, Dell Technologies, and others promote various virtualization or hyperconverged infrastructure environments. They provide the abstraction from hardware and common control planes that operate like clouds and run on commodity hardware and/or in a public cloud.
But there is a sweet spot where IBM is truly differentiated from the rest of the pack: in the domain knowledge of key verticals like financial services, healthcare, or transportation, where it has a long legacy of engagements. It's here where the IBM Cloud itself may play its most prominent role.
IBM's blueprint is vertical industry clouds, equipped with the right frameworks, where IBM can leverage its knowledge, not to mention its longstanding C-level relationships. The vertical industry clouds would be deployed on IBM Cloud, where IBM can ensure the right guardrails are in place. AWS, Azure, and Google Cloud can't really match that, at least not without partners.
That doesn't mean that IBM shouldn't partner in vertical industry clouds. Neither does it rule out IBM playing in the sexier areas like AI where the usual suspects also play, but they are more competitive arenas where IBM and rivals will each have some degree of differentiation.
The Financial Industry Cloud is a clear example of IBM's vertical strategy as it's a sector where it has been well-entrenched. Yes, IBM's getting competition from the usual suspects; for instance, AWS trots out strategic relationships, such as with FINRA, which has committed to a multi-year migration, and Moody's Analytics, which has become a strategic technology partner.
Nonetheless, in sectors like financial services, IBM continues to enjoy a home-court advantage. And so, the Financial Industry Cloud is a walled garden specially set up for the industry's unique requirements for security, audit controls, and regulatory compliance. IBM provides a choice of paths, including lift and shift through VMware, or a transformational approach for refactoring apps into microservices with several services, including managed K8s environment, VPCs, and block and object storage services that have been validated. We expect the list of validated services to grow.
So, at this point in the saga of hybrid cloud, IBM is differentiated in a couple of areas. Its Managed OpenShift service is one of the first alternatives to do-it-yourself that are available in the K8s-anywhere world. But at some point, we expect others to also step up to the plate. But in some vertical industries, IBM does have a leg up that makes its uniqueness more defensible.
We're not saying that IBM should go all out on vertical industry cloud. The company is too large and diversified to paint itself into any single particular corner. Nonetheless, its strongest differentiation for hybrid cloud will be in the familiar terrain where it has lived, cultivated the relationships, and developed the requisite expertise going back decades. And it's a logical place for IBM to amp up its focus.
Disclosure: IBM is a dbInsight client.
Note: We originally omitted mention of IBM Cloud Satellite while reviewing IBM's K8s portfolio, which we have corrected here.