Hyperconverged Infrastructure, Implementation

Hyperconvergence as a Bridge to the Cloud


It's been a little more than 10 years since Amazon launched their Amazon Web Services (AWS) public cloud. As with most new technologies, adoption has taken time; however, analysts predict that by 2020 cloud will be the new normal. This isn't to say that on-premises IT is going away, but things cloudy are clearly the future. So where does hyperconverged infrastructure (HCI) fit in all of this?

The short version is that HCI is a stepping stone between a classic virtualized environment and a full-blown cloud solution. Most organizations can't simply decide "we're going to do everything with clouds now," wave a wand and have it be so. In the real world, organizations of all sizes must deal with technical debts that restrict and restrain them. HCI is the very first step towards getting existing IT under control.

Putting Out Fires

Up to 70% of 308 respondents surveyed by ESG said that they plan to use HCI, in a survey that primarily targeted enterprises. The question systems administrators should be asking is Why? Enterprises can see the writing on the wall. They know the future is cloudy. So why are they investing in HCI?

The answer lies in that discussion about technical debt. Whether their employer is big or small, the number one thing holding most systems administrators back is that the reactive nature of datacenters. You can't automate your environment, research alternative solutions for aged apps or figure out cloud unless you can stop running breathlessly from fire to fire.

Simplifying the underlying infrastructure – from a management perspective, at least – is the first, critical step in evolving away from firefighting. The fewer things an administrator has to manage, the fewer things that can be on fire at any one time.

You can't automate your environment, research alternative solutions for aged apps or figure out cloud unless you can stop running breathlessly from fire to fire.

HCI essentially takes storage, compute (and sometimes networking), and delivers them as a black box. It is essentially buying a bare minimum of IT automation as an appliance. Instead of worrying about partition alignments or resizing LUNs, systems administrators can just stand up workloads and get on with their day. The HCI software handles the problems that live below the virtual machine (VM) so that systems administrators don't have to.

Moving Toward Composable Infrastructure

Once systems administrators have opened up enough time in their schedule to start researching and adopting new IT automation techniques, a cascade effect happens. The more that IT administrators automate, the more time they free up. The more time they free up, the more they can automate.

The ultimate goal for the modern systems administrator is composable infrastructure. This starts with configuration management tools like Puppet, Chef, Saltstack and Ansible, and continues on toward tools like Terraform. Once workloads are composable, they can be easily instantiated on any infrastructure, from HCI to clouds.

While composable infrastructure is the goal, it is something of an unobtainable one; an ideal to strive for, but with the understanding that the end result is always something of a moving target. Each infrastructure is slightly different. HCI and cloud solutions all have different APIs, some support context for VM creation and some don't. The level of support for composability varies from vendor to vendor.

However, the core concept of composability – separating the data being acted upon from the applications (and environments) that act on that data – applies everywhere. And it isn't easy. Making a single application composable can require weeks or even months of research and experimentation. For this reason, HCI's ability to reduce a systems administrator's workload is key.

Getting Your Cloud On

Automation happens a piece at a time. The first step might be getting networking sorted so that DR to cloud is possible. The next might be using templates as the basis for new workloads, and deploying a configuration management tool as part of those templates to ensure that the Operating System Environment (OSE), application configuration and updates are all handled appropriately.

Eventually, automation will turn toward automating the instantiation of VMs, the creation of virtual networks and virtualized network functions, and even the instrumentation of all of the above. HCI solutions that offer access to APIs are a better bridge to the cloud than those that don't, because they allow for more automation. Similarly, HCI offerings that can integrate with composable monitoring applications and other infrastructure automation solutions will provide a smoother transition from traditional infrastructure toward the cloudy future.

None of the HCI offerings available today are perfect in this regard. Each have their strengths and weaknesses. For those considering HCI with an eye toward a future in the cloud, it is worth taking the time to think about what the intermediate steps look like, and picking the vendors that will get you where you want to go.