Are the days where you could stay in high demand by just being the best virtualization administrator over?
Hyper-convergence is having an impact on both the infrastructures you manage and the overall job of the engineer. There's little doubt that service providers will have a steady demand for engineers focused on the compute stack of the data center. The scale of these environments dictates that engineers need to be well versed in the nuances of server virtualization.
On the other hand, hyper-convergence promises reduced operating expenses. To CIOs, this is code for reducing head count. What should you do to keep your skills relevant if your company invests in a hyper-converged infrastructure?
First, it's important to understand the architecture and appeal of a hyper-converged approach. Vendors have attempted to offer converged products for years. HP once offered storage arrays built on their server platform. As x86 hypervisors took hold within the data center, we saw cross-vendor converged products make headway.
Today's hyper-converged options
One of the most popular hyper-converged products today is VCE. VCE's vBlock reference architecture combines Cisco network and server hardware, VMware vSphere and EMC storage. Converged platforms such as vBlock give companies a solid design foundation and one vendor to go to for support. One operational disadvantage to the vBlock approach is the reliance on separate roles or groups needed to managed each sub-component. Most midsize to large businesses still rely on dedicated storage, server and network administrators to manage the converged platform.
Hyper-converged products combine all the major subcomponents in a single appliance or "brick." In most cases, a single 2U server will provide the virtualization platform, memory, CPU and storage needed to run a simple virtualized infrastructure. These products incorporate a scale-out architecture that allows expansion by just adding another appliance to the infrastructure. To this point we've mostly described a combination of commodity hardware, but it's a software layer that makes hyper-converged products truly appealing. By providing a proprietary interface for management, vendors offer a single pane of glass administration for the core infrastructure. This allows a single engineer to manage a large cluster of hyper-converged nodes, and engineers do not have to worry about the complexity of configuring a storage area network. On the low end of the market, vendors such as Scale Computing market their products to organizations without a virtualization admin.
The aim of any hyper-converged solution is to reduce the administrative overhead associated with a virtualized infrastructure, but what skill should you develop to remain a high-demand resource? If you want to continue adding to your technical skill, you can go the DevOps route, or you could explore the business side of IT operations.
Better your technical skills
The drive for operational efficiency will not be solved by convergence alone. Converged platforms don't solve every administrative challenge nor does every workload lend itself to the architecture. The need to automate infrastructure tasks provides the ability to automatically provision servers, VLANs and storage based on a request from the developer or workload demand and will be a critical value-add for enterprise engineers moving forward. Automation tools and cloud management solutions are two exciting technologies that have increasing demands.
To frame the problem, convergence doesn't solve the challenge of requesting and provisioning the infrastructure. While convergence simplifies the provisioning of resources, it doesn't address the challenge of creating VLAN's, making firewall rule changes, creating (or destroying) a virtual machine and all the other tasks associated with actually bringing up a workload. Automation tools include products such as Puppet, Chef and vCloud Automation Center and, generally speaking, take repeatable and programmable tasks and automate them with scripting.
Even if organizations don't have a traditional use case for elastic cloud-aware applications, there's potential value in cloud management products. One of the best use cases is self-service. The ability of development, test and QA teams to provision resources from a self-service portal is a tremendous value. This capability can be found in products such as CloudStack, OpenStack and VMware vCloud Director.
DevOps is one option, but if you'd like to keep a technology focus as a secondary skill, then networking is also a solid option.
Better your business skills
Another way to display versatility is by developing some type of business skill. The two most common skills I see in the field are project management and business analysis. Both roles require an understanding of both information technology and the business supported by the technology. Project management is a fairly common skill to add, as most engineers find themselves in a position where they are managing the implementation of technology projects. For IT pros interested in this route, the Project Management Professional certification allows you to show your knowledge and boost your resume.
Gaining experience as a business analyst can be more difficult. Engineers should seek cross-training opportunities that include shadowing or leading requirements reviews. The ability to translate business requirements into technical requirements shows that you can fully understand the business value of technology. Additional topics that can help in this endeavor include understanding financial terms related to projects and budgets, such as capital expense vs. operational expenses, total cost of ownership, return on investment and net present value.
As hyper-converged infrastructure becomes more popular, maintaining your appeal as a virtualization professional will require adding a skill over and above understanding virtualization technologies.