I recently reviewed IBM’s strategy. It’s a complex issue, which I think can be elucidated by a look at how Red Hat has developed its approach. My Figure above is based on its 2019 roadmap shortly after being acquired by IBM. Open Systems projects are driving the direction of our market, yet the have different values for the code developers and IT operations staff in Cloud Service Providers (CSPs) and enterprises. In particular:
- The largest CSPs (Google, Microsoft, Amazon, IBM, etc.) have the internal expertise to build wide-scale advanced applications without help from outside companies; they participate in Open Source projects to share development costs as they expand into container-based approaches, which has the side-effect of making new technology approaches available to smaller companies (either CSP or enterprise users) to exploit;
- However many (if not most) of these smaller companies lack the technical expertise or time to acquire it;
- Red Hat and a few others make their revenues through deep involvement in and knowledge of Open Source developments to encapsulate new code and techniques to make them easier for enterprises to deploy, as well as providing education and training.
Open Source projects almost always rely on Linux as an operating system, its in-built KVM hypervisor for virtualised computing and Kubernetes for orchestrating container-based applications. Integration of container-based computing into enterprise IT infrastructure is becoming an essential part of enterprise – often referred to as hybrid multi-cloud – computing.
IBM’s new strategy uses OpenShift as the anchor point for its own software and services, which are going through an extensive process of modernization and which started before it acquired Red Hat. The center of gravity for its approach has shifted from z systems and Power hardware as a result, which have become proprietary, but plugged in, ‘bare metal’ elements of this new schema.
The challenge for other systems and cloud suppliers is either:
- To integrate their proprietary hardware (Intel, AMD, Oracle), software (Dell EMC’s Vmware) and services (AWS, Google, etc.) into a wider open source management architecture equivalent of Red Hat’s approach, or
- Concentrate on beating all-comers in providing their (essentially) point solutions, which can be plugged into to someone else’s architecture.
No matter how important Google, Microsoft, Oracle, Vmware or AWS have become to a particular user, they will not want to be isolated from future innovation by the high costs of migration to open source components or solutions.
The challenge for IBM overall is to persuade its customers and prospects that OpenShift is open and cheap enough for them to build their future infrastructure strategy on; for Red Hat – to continue to make OpenShift relevant for non-IBM users to deploy if possible. While OpenShift is no silver bullet for developers, designers and operators of enterprise IT infrastructure, it is currently the most future-proof way to prepare for the transformation in IT ahead.