Businesses today are constantly looking for ways to improve efficiency while saving money – which, if you think about it, is quite an ask. Usually, it takes money to improve efficiency. In IT, this “spend money now to save money later” approach often means looking at new ways of accomplishing existing processes. One of these is hyperconvergence.
What is hyperconvergence? We’ll define it in detail in a moment, but for now we’ll describe it simply as a way of combining and managing the different facets of an IT infrastructure. The goal of hyperconvergence (or hyperconverged infrastructure) is to simplify data center complexity. In turn, this simplifies management and, in theory, saves money.
In this article, we’ll look at the full definition of hyperconvergence and compare it to converged and traditional IT infrastructures. Then we’ll look at the benefits and drawbacks of using a hyperconverged infrastructure. Finally, we’ll consider the use cases for hyperconvergence.
Defining Hyperconvergence (and Other Terms)
In the introduction, we used a very high-level, simplified explanation of hyperconvergence. Let’s look at a more technical definition, courtesy of Wikipedia:
“Hyperconverged infrastructure (HCI) is a software-defined IT infrastructure that virtualizes all of the elements of conventional ‘hardware-defined’ systems. HCI includes, at a minimum, virtualized computing, software-defined storage, and virtualized networking. HCI typically runs on commercial off-the-shelf servers. [..] In HCI both the storage area network and the underlying storage abstractions are implemented virtually in software rather than physically in hardware.[…] Management of all resources can be shared across all instances of a hyper-converged infrastructure.”
For a non-IT professional, this definition may seem a little complicated. But once we understand the basic principles of traditional and converged IT infrastructure, understanding hyperconverged infrastructure is much easier. Before we do that, let’s tackle a question of semantics.
Hyperconverged Infrastructure or Hyperconvergence?
You’ve probably noticed that we’ve used the words “hyperconvergence” and “hyperconverged infrastructure” interchangeably so far. Are they the same thing?
The correct technical term is “hyperconverged infrastructure”, which is frequently abbreviated as HCI. “Hyperconvergence” refers to HCI, but it’s used more by marketers and businesspeople. For our purposes, the terms are indeed interchangeable.
There are some other unfamiliar terms you’ll need to know as we discuss hyperconvergence. They are:
- Hypervisor – A program that creates and runs a Virtual Machine. This is usually done by software, but firmware or hardware can also be used.
- Virtual Machine (VM) – A virtualized version of a computer system; it functions much like a physical computer, but it uses software instead of hardware to run apps and other programs.
- Software-defined storage – Data storage is managed separately or ‘abstracted’ from the physical hardware, allowing a higher degree of flexibility
- Containers – A software package that bundles up everything needed to run the software. This extends from the base code, the settings and all other dependencies. A container is standalone meaning the software will run the same way in different computer environments.
- Virtual Desktop Infrastructure (VDI) – A technology that runs a familiar operating system (like Windows or Mac OS) using virtual machines instead of a desktop computer system.
Now that we know the terminology, let’s move on to the infrastructures involved.
Hyperconverged vs. Converged vs. Traditional Infrastructure
In a traditional IT infrastructure, the computing, networking, and storage components are physically separate and are managed separately. Thus, you’d have computing, storage, and networking hardware; you might even have separate teams to manage each one. The advantage to this is that you can mix and match vendors and alter components as needed, with minimal impact on the other components in the system. The disadvantage is that this setup requires a lot of space and equipment. It can be expensive and time-consuming to implement, monitor, and maintain. And there’s always the issue of capacity; it’s all too easy to under- or over-provide the system.
Converged IT infrastructure attempts to simplify this somewhat by managing the compute, network, and storage layers together while still separating them physically. This has the same pros and cons as the traditional infrastructure, although to a different degree (depending on the actual setup).
In a hyperconverged infrastructure, all three components are bundled together and managed together across all the instances in the system. This can mean using hypervisors, virtual machines, software-defined storage, containers, and other technologies to replicate the functionality of a traditional setup. If you purchase HCI units from a vendor, they come in an all-in-one package. You can also develop your own HCI, although this does require a significant amount of expertise and equipment.
Benefits of Hyperconvergence
Let’s suppose you choose to use an HCI vendor. What are the advantages of this decision?
- The vendor compiles everything into one unit, which saves time and space. It also eliminates the question of integration.
- Datacenter complexity is reduced, especially as compared to a traditional infrastructure.
- Management is done by a single platform for all applications and servers. Capacity, file storage, memory, and network connectivity function like public utilities.
- You can easily add new units when you need to linearly scale your system. (A linear scale is when you need to add the same amount of functionality across all three components – e.g. you need to increase your entire infrastructure by 25% to support a new department.)
- You may need less time and employees to manage the system, but still achieve the same performance.
Drawbacks of Hyperconvergence
Now let’s put the shoe on the other foot. What are the downsides to HCI?
- If you choose to use a vendor, you may be locked into their ecosystem. You’re also at the mercy of their management interface – whatever monitoring and control options it provides is what you can use.
- Not all use cases adapt to linear scaling. For example, you may need to add networking capacity but not storage capacity or vice versa.
- You may not be able to scale individual resources, such as additional computing power. If you need to add more functionality to one resource, you may have to buy an entire new unit, which could be very expensive. It also results in over-provisioning for the non-needed components.
- Combined functionalities mean you can’t fix one component without affecting the others.
When Is Hyperconvergence the Answer?
While HCI seems like it solves the problems of integration, complexity, and scalability, it’s not a universal answer. This is especially true for companies that have unique requirements that can’t be fully met by out-of-the-box solutions.
Where does hyperconvergence shine? The broad answer is when the workload is linear and fairly predictable. For example, adding a bunch of VDIs is usually very scalable in terms of the expected workload. Of course, there are many other potential applications of HCI, including enterprise software, data storage and data analytics systems, web servers, file and print services, and collaboration tools.
