Hyper-Convergence Supercharges Hybrid Clouds

J.P. Doolin


The RightScale “2016 State of the Cloud” survey asked over 1,000 IT professionals, “What is the biggest benefit you are getting from the cloud?” The top response was “faster access to infrastructure,” with “greater scalability” coming in as a close second. More than anything, it seems, businesses in the cloud want agility, and public cloud providers, such as AWS and Microsoft, have changed the measuring stick. Customers want something that is faster and easier to use – point, click and “ding,” your server is ready – and now they know they can get it.

As the public cloud has upped the competitive ante for hardware vendors, many infrastructure companies are working hard to show customers that they’re listening. The first wave of this effort focused on converged systems, in which separate components from different vendors were stitched together and tested for the customer, increasing speed to deployment from months to weeks. These systems were successfully released before the public cloud had gone mainstream and include Flexpod (Cisco/Netapp) and vBlock (Cisco/EMC/VMware).

Now we’re seeing hyper-convergence go beyond early adopters and become mainstream, which, like the public cloud, provides enterprises with faster access to infrastructure, to a point where the underlying hardware doesn’t matter anymore. SimpliVity and Nutanix were the first hyper-convergence startups, and Cisco’s HyperFlex and the HP’s hyper-converged platform have recently entered the market.

These companies give their customers cloud-like agility to deploy IT infrastructure. Individual nodes (compute, memory, storage, networking) have all the resources and intelligence to run applications that can be connected together seamlessly to provide scalability and resiliency from failure. Distributed brains that all work as one – genius! Installation is 30 minutes, and adding nodes to an existing cluster is software defined and can happen in seconds. I predict within the next three years, new private cloud infrastructures will be comprised of mainly hyper-converged systems, based on their ability to integrate with the public cloud.

So where does this leave the customer? Hyper-convergence has nearly matched the agility and scalability of the public cloud, so the decision becomes more of a business one: Should you own or rent your infrastructure? At Equinix, we see nearly all of our Fortune 5000 customers choosing both. Being able to deploy workloads in either fashion and move them back and forth is where everyone wants to be – this is a true hybrid cloud.

For example, if you have a production environment where the compute, memory, and storage resources are known and their growth predictable, then owning (private cloud) makes financial sense. However, when you temporarily need a burst of resources, such as for seasonal demand or to spin up an application for testing and development, then renting those resources in the public cloud is a good option.

What sometimes gets lost in the dialogue between public and private clouds is the interconnection between them and ensuring that it does not pose a bottleneck. This is where Equinix comes in: We enable high-bandwidth, low-latency connections that seamlessly integrate virtual and physical worlds.

Within this emerging landscape, cloud leaders such as AWS, Microsoft Azure/Office 365, Google Cloud Platform, SoftLayer and Oracle Cloud are turning to Equinix to make their access points more globally available and connect them to the outside world. Our enterprise customers deploy private cloud infrastructures in Equinix data centers for the best possible connection to the public cloud. In turn, Equinix ensures the connection between private and public cloud is secure, direct, and in the closest proximity to their public cloud workloads for the best possible performance and less risk. And because we’re vendor-agnostic, enterprises are free to choose the path that best matches their business, workload and budget requirements ̶ be that hyper-convergence, public cloud or both.