Five things I learned about hybrid cloud from IBM clients
Face-to-face conversations with clients provide a great opportunity to hear first-hand what challenges business leaders are facing around hybrid cloud for enterprise workloads. Whether these conversations take place at a conference or onsite during a client project, they’ve shown me what’s right and what’s wrong with cloud from the perspective of the IT administrators whose job it is to keep manufacturing floors, distribution centers, branch offices, pharmacies and a host of other businesses running. What follows are some of the key takeaways I’ve gleaned from recent discussions with clients on enterprise hybrid cloud:
1. Hybrid cloud is real
The same admins who in recent years kept saying, “No, the public cloud is something our web/mail/application team is doing,” are now saying, “I really need to look at disaster recovery in the cloud for these core-business workloads.” Or, “The production databases will stay on premises, but I need to give database admins the ability to provision test environments in the cloud.” And while we’re talking about public cloud, they often remind me that they still have the on-premises environment to manage, and too many tools is a bad thing. IBM’s focus on tools to manage across the hybrid cloud seems to be making heads nod. Because hybrid cloud solves real business problems for enterprise workloads, and because cloud management solutions are now readily available, hybrid cloud is turning from “distant possibility when I have time” to “urgent project I need to undertake in the next six months.”
2. Self-service works better on public cloud
I’ve been on a journey with many clients to build a private cloud and deploy their core business workloads quickly and seamlessly across IT siloes. And a big percentage of them have completed that step of the journey. But private cloud provisioning of enterprise workloads is an admin’s game. In 99 percent of cases these admins will tell you, “There’s no chance we’re letting end users directly access these cloud tools and deploy their own.” It’s a matter of security, compliance and proper allocation of a limited pool of resources. However, giving these same end users — usually we’re talking about app developers — the ability to deploy their own environments in the public cloud: now that alleviates a lot of concerns about access, ownership and proper billing for use.
3. Cloud-native workloads are gaining speed in traditional IT
When we say “cloud-native” in this context, we mean apps running in containers, as opposed to whole VMs. Until mid-2019, this topic also seemed to fall in the category of, “This is outside my purview and interest as a core-business admin. We have another team looking at containers.” But I’ve now heard from a number of clients that use cases for cloud-native apps, based on real business needs, are emerging. Whether it’s a Node.js application running a company website or a development pipeline with Git and Jenkins, organizations are finding themselves needing to stand up containers side-by-side with their core-business apps in VMs. And needing the tools to manage across VMs and containers.
4. Not everyone has embraced private cloud yet
With some exceptions, usually for performance, most enterprise workloads have been fully virtualized for close to a decade. But even as we discuss public and hybrid cloud more and more, some businesses haven’t yet taken the big initial step of starting with a private cloud. What’s holding them back? For some, it’s the firm boundaries among the compute, storage and network IT siloes. While the boundaries are there for a reason — guaranteeing reliability and compliance — they continue to hold admins back from using any private cloud tools that reach across the siloes. For others, it’s the simple lack of system and human resources to initiate a proof-of-concept, test out the tools and convince management of their business value. In those cases, the option to prove out running core-business workloads in the public cloud might be the easier start on the cloud journey.
5. Package automation convergence is a necessity
Clients from every industry have raised the same challenge: it simply will not do anymore to have different tools for managing software package deployment, versioning and compliance across different platforms. Whether it’s through Ansible, Terraform, Chef or Puppet (or some other local favorite), package automation must extend across IT tiers and types of workloads and use a single tool. In most cases, the automation tool is already well in production and simply needs to be extended to core-business environments. IBM has completed many such projects for enterprise clients, unifying architectures and operating systems under a single file set automation solution.
What I’ve heard from IT administrators illustrates the resource and process challenges admins face when going on the cloud journey. Not only that, but I learned that IBM’s approach to hybrid multicloud for enterprise workloads is spot on in addressing the real needs organizations today face.