Most cloud vendors are obsessed with moving clients to their platforms. They understand that although their core services, such as compute, are no longer differentiated, there is still a lot of money to be made just by hosting their clients’ workloads. And they realize that this migration madness has sufficient legs to last for at least three to five years. No wonder migration spend constitutes 50-60 percent of services spend on cloud.
What about the workloads?
The cloud destination – meaning the platform it operates on – does not help a workload to run better. To get modernized and even natively built, a workload needs open architecture underpinned by software that helps applications be built just one time and run on any platform. Kubernetes has become a standard way of building such applications. Today, every major vendor has its own Kubernetes services such as Amazon EKS, Azure Kubernetes Services, Google Kubernetes Engine, IBM Cloud Kubernetes Service, and Oracle Container Engine for Kubernetes.
However, Kubernetes does not create an open and portable application. They help in running containerized workloads. But if those workloads are not portable, Kubernetes services defeat the purpose. Containerized workloads need to be architected to ensure user and kernel space are well designed. This is relevant for both new and modernized workloads. For portability, the system architecture needs to be built on a layer that is open and portable. Without this, the entire migration madness will simply add one more layer to enterprise complexity. Interestingly, any cloud vendor that helps clients build such workloads will almost always be the preferred destination platform, as clients benefit from combining the architecture, build, and run environment.
Role of partners of cloud vendors
Cloud vendors need to work with partners to help clients build new generation workloads that are easily movable and platform independent. The partners include multiple entities such as ISVs and service providers.
ISVs need to embed such open and interoperable elements into their solution so that their software can run on any platform. Service providers need to engage with clients and cloud vendors to build and modernize workloads by using open, interoperable principles. As there is significant business in cloud migration, there is a risk that the service partners will get blinded and become more focused on the growth of their own cloud services business than on driving value for the client. This is a short-term strategy that can help service providers meet their targets. But it will not make the service provider a client’s strategic long-term partner.
Role of enterprises
There is significant pressure on enterprise technology teams to migrate workloads to the cloud. Many of the clients we work with take pride in telling us they will move more than 80 percent of their workloads to the cloud in the next two to three years. There is limited deliberation on which workloads need newer build on portable middleware, or even if they need a runtime that can support an open and flexible architecture. Unfortunately, many enterprise executives have to show progress in cloud adoption. And though enterprise architects and engineering teams do come together to evaluate how a workload needs to be moved to the cloud, there is little discussion on building an open architecture for these workloads. A bright spot is that there seem to be good architectural discussions around newer workloads.
Enterprises will soon realize they are hitting the wall of cloud value because they did not meaningfully invest in building a stronger architecture. Their focus in moving their workloads to their primary cloud vendor is overshadowing all the other initiatives they must undertake to make their cloud journey a success.
Do you focus on the architecture or the destination for your workloads? I’d enjoy hearing your experiences and thoughts at [email protected].