Hyper-converged infrastructure is one of the fastest-growing trends taking over data centers today. Many organizations have either already moved or heavily considering moving to this architecture. So, what’s the big deal about hyper-converged infrastructure? Let’s begin with an explanation of the traditional data center design and what differentiates hyper-converged infrastructure from it.
Many know the traditional data center as the server, storage, and virtualization stack. By that I mean that it has a disparate inventory of Intel-based servers, a.k.a compute nodes, with a dense composition of memory, CPU cores, and network interface ports. These compute nodes will probably have a no-disk footprint, or just enough disk space to install a very small image of a virtualization operating system (e.g., hypervisors like VMWare ESXi or Microsoft Hyper-V). At the bottom of the stack, you will find one or more humongous disk storage arrays, which is the repository of all the data, virtual machine images, and perhaps backups.
Many data centers have several different storage arrays with specific use cases or workloads. Storage arrays have built-in controller systems (another set of embedded Intel-based servers) that are purpose-built for managing and provisioning logical capacity, I/O, performance, and disk services. Another set of devices called fabric switches connect to the compute nodes and communicate to the front-end storage array controllers to ensure high availability and shared storage capabilities in a shared or clustered format.
That’s just the physical part. The logical and virtual architecture requires software and cluster management to ensure that all these physical devices interact with each other harmoniously in a complex data center environment. Mind you, this is just the tip of the iceberg. The rest of the functions and features of this stack would require a much longer article, and we are only talking about a single data center or a single stack. Some organizations have multiple stacks across several data centers doing either redundant or unique things.
The traditional architecture is somewhat hardware-centric. This means that in order to manage, operate, administer, scale/upgrade, and troubleshoot, you must touch each of these disparate systems while also making sense of it as a whole. Don’t get me wrong, this architecture is a proven design. It exists (and has existed for many decades) in everything from the smallest data closets to the largest data centers. Many will keep maintaining and renewing the components to preserve this type of a stack. Hardware manufacturers continue to produce devices specifically for this design, and the likelihood of it going away soon is slim.
Now that you’ve had a crash course on the traditional architecture, let’s dive into hyper-converged infrastructure, which is also referred to as a software-defined data center (SDDC). Unlike the traditional hardware-defined data center, in this data center, the compute nodes, storage array, storage fabric, and virtualization stack are all rolled up in a hyper-converged node or appliance.
Essentially, it’s a server CPU/memory with built-in storage and a disk controller installed with a hypervisor and hosted virtual machines and data. So why all the hoopla and fancy terms when it’s only a standalone server? Well, it’s actually a lot more than that. In a software-defined, hyper-converged architecture, all those physical devices (especially the storage and fabric) are abstracted from hardware and presented to the virtualization stack like it is a legacy traditional design. Hence the term software-defined data center. The aggregation of these components in a redundant or N+1 (or more) system makes up the hyper-converged infrastructure. The controllers, fabric, and virtualization are all running as modules or virtual components in the hyper-converged nodes. Should a node fail and take down access to the compute and storage, the redundancy of the nodes is designed to allow the data center to sustain operations and availability.
How can you build a business case to shift to hyper-converged infrastructure? Here are a few great reasons for making the switch:
I can go on and on describing how great hyper-converged infrastructure is, but the more important determination is whether or not it is a good fit for your organization. Examine the reasons above. Do they align with your organization’s goals and objectives? If the answer is yes, it’s time to begin the process of moving to the modern data center.
ConvergeOne's Data Center Experts have written a guide that provides valuable insights about how you can strategically design your data center infrastructure to power the technology of tomorrow. The following areas are explored:
Digital Transformation
Data Protection
Active/Active Data Centers
Converged Architecture
Hyper-Converged Infrastructure
Application Centric Infrastructure (ACI)