Creativa - Fotolia
For those virtualization admins hiding under a virtual rock regarding cloud, I have news for you. Your job isn't safe. No one can put the cloud genie back in the bottle. Cloud computing is here to stay, and virtualization admins need to shift focus to keep up with tomorrow's jobs.
The move to cloud is already happening at all levels, from the smallest through to the largest businesses. Cloud and microservices mark a new iteration of change that is as disruptive as the original arrival of virtualization with VMware -- if not more so.
What we're good at: Hardware utilization and availability, cost savings
Virtualization has two phases: consolidation and abstraction.
In the beginning, virtualization's goal was more efficient use of underutilized hardware. Rarely do servers consume all the resources allocated to them. Virtualization admins could reclaim these lost resources and vastly reduce wasted capacity.
In phase two, virtualization developed advanced functions such as live storage motion or migration, high availability and fault tolerance. These virtualization capabilities address the issues that arise when several machines share one physical piece of hardware. Automation arrived and made server deployment simple and straightforward.
I argue that this virtualization adoption curve peaked a few years ago -- we are now moving to the next iteration, and you'll need to follow a cloud computing career path to come along.
Where we're going: Cloud-based services, no hardware costs
Even once-conservative technology adopters, such as financial institutions, are jumping on board with the third wave of virtualization.
There is a thirst to cut costs, and automation allows massive cost cuts. There will be job losses. No virtualization admin should think it will never happen to them. You are fooling yourself. Fewer staff means fewer medical plans and pensions to support. It is not hard to see why the cloud appeals to the bottom line.
There will not be enough cloud computing careers to go around based on old virtualization working practices, such as in a phase one scenario.
Consider virtual machine orchestration. In early-phase virtualization environments, VMs still required some level of administration action, such as deployment from a template, to accompany automated steps. Tools such as VMware vRealize Automation or Platform9's managed vSphere stack enable an approved user to request a VM, customized to their specifications, and have it deployed within 10 minutes with no administrator interactions. Larger companies used to have several virtualization admins whose jobs purely entailed VM creation and deployment. Within a year or two, that job role disappeared.
Why the virtualization admin is a dying breed
Virtual machines are now moving to cattle status, meaning they're disposable commodities. To scale applications, organizations adopt automation tools that deploy new VMs. It's quicker to deploy another instance of a machine than to troubleshoot a broken one.
DevOps does away with manual work; manual deployment is the exact opposite of how DevOps is supposed to work. A key tenet of DevOps is that tasks performed more than once in the same way should be scripted so the IT platform does the action itself.
Platform as a service reduces the workload. Workloads that used to be custom-built and based on infrastructure as a service are now provided as a service for consumption by developers and companies. Examples include the larger cloud vendors offering secure and highly available database hosting that organizations consume without any effort to build and manage the underlying database infrastructure. Little to no database admin input required. No server admin required either.
The complexity hasn't gone away -- it has just changed. Management complexity moved from the VMs to orchestration and scaling. Virtualization elements such as high availability and disaster recovery (DR) lost importance, while the IT industry turned its attention to microservices that are scalable, redundant and can be spun up and down at will. Automation means little to no hands-on intervention. For example, you can spin up a cloud infrastructure from a single PowerShell script.
Classic DR locations are now costly relics of waste. Cloud affects virtualization in secondary ways. For example, businesses are used to having one primary data center and one DR setup in another data center. Given a relatively modern application set, the entire company infrastructure can restart in its entirety in the cloud in the event of a disaster. Modern DR management products, such as Zerto and Acronis, eliminate the costly secondary data center, allowing businesses to prepopulate and configure DR setups in the cloud.
This is the reality for virtualization admins, and the only future is in a cloud computing career. Over time, more applications are built cloud-first to save money from the start; the old, immovable on-site applications go the way of pagers and typewriters.
The reality is that most virtualization admin roles as we know them will vastly shrink or become outmoded over the next decade. A virtual data center requires far fewer staff, and with automation and scripting, a single administrator can manage massive numbers of servers.
Cloud computing careers for virt veterans
There is still time to retool and get on a cloud computing career path. Virtualization admins are luckier than most. While the technology itself may change, these administrators have skills that easily translate to the popular cloud and DevOps arena.
This doesn't mean becoming a code guru or programmer, but a virtualization admin will need a deep understanding of architectures and tools such as Docker for containerization, Chef for configuration management and Kubernetes for container orchestration to become a DevOps admin. Learn multiple scripting languages and investigate hyper-converged infrastructure for cloud hosting.
The warning signs are there, fellow admins. It is just a case of doing something about it while you still can.
Help keep an organization on track as a cloud capacity manager
Break down seemingly convoluted DevOps job requirements
DevOps engineers must demonstrate strong communication skills
Set up a DevOps home lab to gain hands-on skills