A titanic battle for control over the future of enterprise has begun in earnest with the rise of distributed cloud computing architectures through which IT organizations will one day centrally manage a wide range of application environments.
All the major cloud service providers are already extending the reach of their platforms with that goal in mind. Google, for example, has created the Anthos platform based on Kubernetes that can be deployed on any cloud or on-premises IT environment. IT teams can then centrally manage workloads as they best see fit regardless of where they are running.
Microsoft, meanwhile, has launched Azure Arc, a management platform that makes it easier to deploy and manage Azure services across multiple clouds and on-premises IT environments. Each resource is assigned a unique Azure Resource Manager ID that enables it to participate in a resource group and be assigned tags like any other Azure resource.
Amazon Web Services (AWS), meanwhile has developed AWS Outposts, a managed service through which instances of its operating environment can be deployed in an on-premises IT environment or at the network edge. AWS has even gone so far as to build its own servers for those environments.
Providers of on-premises IT platforms are not simply rolling over as cloud service providers attempt to lay claim to their turf. Both Dell Technologies and Hewlett-Packard Enterprise (HPE) are extending managed services they provide to the cloud as part of an effort to make it possible to manage an extended enterprise via a console they provide.
IBM, meanwhile, sees an opportunity to regain supremacy by enabling a hybrid cloud computing platform via a Red Hat OpenShift platform based on Kubernetes that can be deployed anywhere. IBM Cloud Satellite is a managed service that extends the reach of IBM’s ability to centrally manage multiple clouds all the way out to the network edge.
Consoles and Crossplane
The one thing all these variations of a distributed cloud have in common is that they assume a console accessed via proprietary service is at the center of an extended enterprise. However, that may not necessarily be how distributed clouds ultimately manage themselves. An open source Crossplane project being advanced by the Cloud Native Computing Foundation (CNCF) is leveraging the control plane originally developed for Kubernetes to create a framework for managing IT resources on any type regardless of location.
Crossplane, rather than limiting the ability of that control plane to manage Kubernetes clusters, makes it possible to also orchestrate legacy virtual machine environments using the Kubernetes application programming interface (API). Originally developed by Upbound, the company recently unveiled a managed service dubbed Upbound Cloud through which it provides access to a curated instance of Crossplane. However, there’s nothing stopping either an internal IT team from deploying Crossplane themselves or contracting an IT services firm to deploy Crossplane on their behalf.
“Enterprises want to own the control plane for all the different clouds,” says Upbound CEO Bassam Tabbara.
Managing IT Costs
In addition to not wanting to become locked into a single platform, IT teams are under increased pressure to reduce the total cost of IT. Each platform added to an IT environment requires separate tools to manage it that someone in the IT organizations needs to learn how to use. Before too long there is now a separate team of specialists that has been hired to manage each platform.
Distributed cloud computing environments in theory present an opportunity to centralize the management of IT in as much a cloud computing platform can be stood up. Not every class of edge computing platform has the compute and memory resources needed to run a full stack of cloud software. In fact, Gartner predicts that by the end of 2023 only 20% of edge computing platforms will be delivered and managed by hyperscale cloud providers.
One way or another, however, the way enterprise IT is managed is about to fundamentally change. The existing tools IT teams rely on are largely designed for on-premises IT environments. IT teams — either through new tools accessed via a console provided by an IT vendor or one they construct themselves — are required to manage distributed computing environments made up of virtual machines, bare-metal servers, graphical processor units (GPUs), containers, Kubernetes clusters and serverless computing frameworks. Achieving that goal will require substantial investments in, for example, modern automation frameworks that enable IT teams to manage infrastructure as code.
IT vendors are making a case for providing access to those tools via services they provide. Each organization will need to decide for themselves to what degree that approach makes sense for them. Some organizations, for example, may decide to focus their limited resources on building applications rather than managing the infrastructure they run on. Others will decide IT is still too crucial to trust the management of it to anyone else. No matter the path chosen, the one thing that is clear is distributed computing is about to be taken to a whole new level.