Hyperconverged infrastructure (HCI) appears to have been confusing people ever since 2012, when the term was first coined by Steve Chambers and Forrester Research. While computing experts agree on the basics of HCI as a system of IT that’s intended to bring a more powerful virtualized infrastructure to big enterprise, there is still a lot of disagreement as to what defines and delineates HCI.
By the standard definition, in HCI the elements of computing, SAN, and networking are virtualized and all the networking and storage tasks are implemented virtually using software rather than physically in hardware. But since “hyperconverged infrastructure” originated in marketing lingo rather than in purely technical terminology, its meaning isn’t quite so defined. Divergence of opinion about what constitutes HCI shows that even experts are uncertain about the role and importance of the approach.
How are we to understand what HCI is and the role it’s likely to play in enterprise IT? We decided to go straight to the top to get a clear answer for you. We asked three experts with long-term experience in data center and enterprise IT how they define hyperconverged infrastructure, and this is what they told us.
So What Is Hyperconverged Infrastructure?
The three experts we turned to for answers about HCI and its value to in the future were Maish Saidel-Keesing, Petar Marinkovic, and Aleksandar Nenov. Below we get a chance to find out more about each them, their backgrounds in enterprise IT, and just what they think HCI is all about.
Maish Saidel-Keesing is a cloud architect and virtualization veteran who has been involved in IT for the past 17 years. He is a world-renowned author of two books on enterprise architecture: VMware vSphere Design Guide and OpenStack Architecture Design Guide. He has a popular blog Technodrone and most of the time he can be found lurking on Twitter (@maishsk).
Maish: To me, Hyperconverged infrastructure is an all-in-one, tightly integrated solution for network, storage, and computing that enables you run anything and everything you need in one place. HCI will usually be provided by a single vendor. There are a number of vendors that provide HCI solutions. VCE used to dominate the market with the vBlock, and today Nutanix probably has the most compelling solution.
Small form factor also plays a very important part in the HCI solution. The ability to manage everything under a single user interface—and even more importantly, under a single API—is crucial. The important thing to understand about HCI is that it is not a solution in itself, but a means to an end. It allows you to deploy, operate, and in some cases even to scale your applications in a highly optimized environment with zero to no latency, which will suit certain application workloads and solutions. The application layer on HCI needs to be able to make full use of extreme proximity to the storage and network resources in order to provide a streamlined and optimized solution.
However, I personally think that HCI will not last. The reason being that the only players that use HCI in the market today are enterprise companies who use on-premises deployments. Many organizations are moving to the cloud to address their ever growing compute needs and leaving the on-premises world. As this trend continues only major companies will pursue a hyperconverged infrastructure solution. The big cloud providers have gone past what HCI can offer and are already designing their own hardware to suit their explicit and exact needs. They will not be needing third-party vendors to provide a solution for their huge scale of operations.
Petar Marinkovic is an IT professional with more than 11 years of experience with systems, network, virtualization, and, in the last few years, the cloud, primarily with a focus on AWS. Petar sees his work with the cloud as a perfect synergy of all of the IT experience he has had so far, and he has high hopes that in the future, IT will be more simplified and automated. On the tech side, his other passions are software development (Petar started his career as a Java developer) and writing (which is still in the very early stages).
Petar: Hyperconverged infrastructure, in my opinion, is your complete enterprise infrastructure offered as a one product, controlled under one interface (GUI or CLI) and usually provided by one vendor. HCI provides a similar solution to what you get in cloud, where public cloud providers offer compute, storage and network services, but on your own premises. Nutanix is doing quite a good job recently with their vision of how HCI should operate, and I think they’re doing a good job of pushing this trend forward.
Personally, I have never liked the idea of being bound just to one vendor. Migrations are a pain, you need to have dedicated staff just to support one technology stack, and you’re pretty much locked in to only one way of utilizing your on-premises infrastructure. With so many enterprises moving to cloud, either as a DR solution or completely shifting their operations from on-premises data centers, I don’t know if hyperconvergence can be a competing technology at mass scale. There will always be businesses that won’t move to the cloud, but it’s highly doubtful that HCI companies can win those enterprises over to replace their current on-premises setups with a hyperconverged approach.
Aleksandar Nenov is a senior IT professional in cloud web operations and Managed AWS Services, with a deep understanding of AWS solution approaches from a business, technical, and service management perspective. Since 2009, he has helped dozens of small and large businesses benefit from running their workloads in the cloud. Aleksandar is one of 35 global AWS Community Heroes and through his project AWS User Groups Support (@awsugsupport on Twitter) he helps local AWS communities throughout the world.
Aleksandar: HCI systems can be defined as a platform for deploying on-premises infrastructure, where its core elements are integrated compute, network, and storage infrastructure along with a software above the virtualized components which have control over the complete system. This software-defined storage, networking, and compute allows easy access to resources through API calls for on-premises environments.
Research company Forrester predicts that HCI systems will become prevalent as a platform for deploying on-premises infrastructure. On the other hand, big cloud vendors are already designing hardware on their own platforms to support resource access through API calls.
Even though Forester firm predicts that HCI will become ever-present as a common platform for deploying on-premises infrastructure, a vast number of organisations is leaving the on-premises world, choosing the cloud as a solution for their growing compute needs.
The Final Word on HCI
In Stratoscale’s view, hyperconverged infrastructure is a useful building block. HCI can be utilized as a way to pool together resources so as to maximize the interoperability of on-premises infrastructure. In many ways, HCI is seen as a way of creating a private, on-premises interconnected cloud architectures. But while HCI could continue to be a functional approach for some large companies and certain types of applications, we don’t see it as a long-term solution for big enterprise. The real future in enterprise computing is full or hybrid deployment in the cloud.
HCI is not a truly equal alternative to hybrid or private cloud infrastructure. HCI is frequently not optimized for networking. Third-party elements need to be bolted onto the infrastructure and configured separately, often causing slower actualization times for apps. HCI is also more complicated to service, requiring a dedicated IT team at a time when most companies are looking to deploy a slimmer and more agile IT workforce. Other private on-premises cloud capabilities such as multi-tenancy, automation and cost-effective scaling are not fully available to HCI customers. A truly hybrid cloud infrastructure enables companies to manage and move workloads between clouds, permitting all users to utilize the same methodologies.
So now that you’ve heard from the experts, it’s up to you to decide. Do you see the potential hyperconvergence has to integrate all of your disparate on-premises deployment components, or do you think that you’d rather avoid the risk of building your company’s infrastructure around a model which may or may not be effective in the long run when it comes to doing business in the cloud? What’s your opinion about the role of hyperconverged infrastructure and its value in large enterprises today?