Converged infrastructure combines compute, storage and networking technologies for data centers. Although the converged components are all housed together, they each retain their independent functional identities. However, vendors pay close attention to enhanced connectivity by validating compatibility before point of purchase. In short, everything works together right out of the box, either as a connected unit or separate modules.
Hyperconverged infrastructure shares some of the benefits of converged, but has its own advantages and unique design. With hyperconverged, the architecture is a software-based solution running on “white box” hardware. Hyperconverged compute, storage and networking capabilities exist virtually, making them inextricable from each other.
So, now that we’ve got a rough outline of the similarities and differences, let’s compare how converged and hyperconverged infrastructures stack up.
Round One: Physical Footprint
First, in deciding which of the two architectures would potentially deliver the greatest operational benefits, data center managers must decide whether size matters to their enterprises.
Hyperconverged is a single platform, and all-inclusive. Converged is the best three products, compute, storage and networking, linked together to create an ultimate platform. In most cases, hyperconverged offers a smaller physical footprint than traditional converged infrastructures, no matter what components go into converged. Because hyperconverged relies more on software over hardware, it takes up less room, making it ideal for small remote data center locations where spatial economy reigns supreme.
Round Two: Scalability and Modularity
According to analysis from Research and Markets, the data center construction industry expects to expand by more than 9 percent annually until 2019. Data center customers are a driving force behind these increases, as service providers push to offer wider product portfolios, as well as collect and analyze more data.
That said, scaling up data centers can’t be done indiscriminately, otherwise data centers could suffer for it. Converged infrastructure allows data center managers the freedom to expand certain components in their architectures depending on their needs, and leave others alone, thereby minimizing the potential for waste.
Imagine you’re a CIO, and you’ve got a large converged installation. Let’s say you’re adding a new application to your environment that requires you to store images. You don’t need more compute power, just more disc space. In a converged environment, you’d simply grow the storage aspect and the compute resources would stay the same.
Since hyperconverged comes in a single platform, it’s impossible to add any one component without purchasing a whole other module.
Round Three: Energy Consumption
It’s not a secret how badly data centers must monitor electricity use, from a cost basis and for the sake of the environment. So, is a converged architecture more costly than hyperconverged, or the other way around? It depends on your perspective.
If we’re talking very small platforms, hyperconverged will usually be more cost-efficient in terms of power. Even a single converged module still needs network switches, storage controllers, discs and servers. However, as a data center with converged infrastructure scales up and utilizes the technology’s modularity, it could save more power than a hyperconverged architecture going through a similar growth spurt.
In terms of which is the “best solution,” it strictly depends on your data center environment and the overall goals of your organization. If you’re looking for scalability, traditional converged infrastructure would be your best bet. But if you’re looking to minimize your physical footprint, you may want to look into hyperconverged. Both of these architectures have clear strengths and weaknesses, so it’s crucial you know exactly what you’re looking for before you upgrade your data center.
For more information on converged infrastructure, please read our Red River ViewPoint: https://www.redriver.com/wp-content/uploads/2015/09/rrv-viewpoint-Converged-Infrastructure-FlexPod.pdf
About the Author:
Mickey Cullop is a Sr. Systems Engineer and confirmed Data Center Evangelist for Red River, a leading technology integrator that delivers best-in-class technology solutions and services to military, intelligence and civilian agencies and the companies that serve them. Mickey is the resident expert on the various Red River hardware applications with a true passion for discussing new technology trends and advancements.