Within the last year, one of the most common questions I’ve been asked by our customers is “should I introduce hyperconverged infrastructure (HCI) into my data center architecture?”
This is a valid question. HCI is a shiny new solution on the tech shelf and based on its "all-in-one" cost-effective approach to storage, compute, networking and management, business leaders and IT teams are ready to buy.
HCI is an infrastructure system with a software-centric architecture that integrates compute, storage, and networking. Since HCI pulls these resources into one appliance from one vendor, HCI is commonly referred to as an “all-in-one” approach, and this simplified approach is a main factor in its rapid market adoption and growth.
The first wave of HCI solutions in the market tried to fulfill this need but weren’t as robust as some of the newer solutions. They lacked in performance, scale, and reliability which made it difficult for customers to realise the full potential of a truly hyperconverged solution. However, things are now changing fast. Before we get there, let us think what are the key problems HCI is helping solve.
What does HCI help solve and why is it fueling so much interest?
There are a few key reasons why HCI is climbing up the CIO’s most wanted charts.
Many of my CIO friends in the industry tell me that they don’t have a budget to staff a network administrator, a storage administrator, an application administrator and so on. These are the roles that you could still see in a traditional data center. The CIOs are under huge pressure to reduce costs.
Today most of them say- I want to receive a box and that box should have very few dials that a generic administrator can turn as easily as possible and get the work done. As a fallout, the specialist skill sets are eroding making way for the generalist administrator. So “Simplicity” is the first and biggest demand of today’s complex IT landscape.
Reason number two and three lie in the workloads and Quality of Service (QoS). Fundamentally people thought that HCI is for one purpose and that is to put an application on it and run it. The workload is the amount of processing that a computer system must complete- including an application as well as the users who may be interacting with the application. It is also the scale of the application to serve as many users at a given point in time.
“According to a recent Gartner forecast, HCI revenue is expected to grow at a 48% annual growth rate through to 2021, to reach $10.8 billion.”
But a number of CIOs were not excited about being pigeon holed to a single workload on HCI – as some of the first generation HCI providers operate. The next gen of HCI is what they are looking for. Being able to provide an HCI environment that does not restrict users to single workloads is the call of the hour. This is especially true as workloads that requires a cloud experience become a reality. Reliably in a multi workload environment is a big question mark for the older gen HCIs. Ergo, workloads are the gold standard of defining HCI and what it does.
The QoS is a critical enabling technology to deliver consistent performance to business-critical applications. QoS is essential for systems running varying workloads with unpredictable demands, such as databases, virtual machines, and private clouds. Better QoS means better control of every application, and, in turn, increased operational efficiency and maximum resource utilization. Today, you can actually deploy multiple workloads and still be able to manage cost quality and service with some guarantees on performance and scale out.
That brings us to Scale- which is the fourth pillar of HCI. Say you have one box and you are reaching a limit on the applications you can run on it. Or say you are using an application but maxing out on data. Today’s HCI solutions offer a much more flexible scale out feature-set with the ability to expand either of trinity of compute, storage, and networking - irrespective of the other two. Which means you can actually scale out independently on say networking, but you don't have to scale on both compute and storage!
The bottom line
The hype for hyperconverged is real. However, making business decisions based on trends or a fear that you will miss out is not a tactic I would endorse. In most cases, and especially for HCI, it pays to do your homework.
There is no short answer to determining whether or not HCI is the right choice, but a good place to start is by developing a deeper understanding of your business needs. To begin, survey your infrastructure, map out what you want to accomplish by migrating to hyperconverged, connect with potential vendors and, of course, examine your workloads.
Deepak Visweswaraiah is senior vice president & managing director, NetApp India.
Disclaimer: This article is published as part of the IDG Contributor Network. The views expressed in this article are solely those of the contributing authors and not of IDG Media and its editor(s).