Building Your Own Hybrid Cloud – What is Hyperconvergence?
Posted May 15, 2019 Justin McDonald
Customers for decades have been relying on what is commonly described as a 3-tier data center “server-storage fabric-storage platform” architectures to provide application services to their internal users and external clients. These solutions continue to evolve today with faster and denser solutions, but due to the likelihood of multi-vendor solutions involved in this “stack”, it is not as manageable as any IT professional would like, “vendor finger pointing” can occur in mixed environments, and scaling for growth can be both unpredictable and expensive. Below is a diagram (Figure 1) of a generic, complex 3-tier infrastructure that is being described, along with the virtualization layer discussed below:
Virtualization technologies such as VMware vSphere, Microsoft Hyper-V, and KVM have allowed for organizations to do more with their server investments (moving from pre-virtualization utilization average CPU utilization of below 10% to 50% or more depending on the design of the virtual cluster); however that has also necessitated the need for more advanced and higher performing shared storage platforms and while virtualization allowed for the data center footprint to shrink, the manageability of the equipment and virtual machines has remained a challenge. Virtualization has also led to what is coined “VM-Sprawl”, which means that in many cases more virtual machines were created and then not destroyed when no longer required, taking up valuable storage space, using 3rd party licenses, etc.
Hyperconvergence or hyper-converged infrastructure (HCI) collapses the stack into an appliance-based solution, reducing data center footprint further, removing or abstracting the difficulty of separate server, storage, and in some cases, hypervisor/virtualization management into a single frame (these can vary from use case to use case from small 1U appliances for remote office/branch office (ROBO) needs, to traditional 2U rackmount appliances, and in some cases which have specialized mass-storage needs, 4U appliances can be used. An example of a hyperconverged infrastructure from a logical diagram perspective is show in Figure 2 above; an additional item needed to complete an HCI solution would be top of rack switches, generally 10GbE is the current recommendation for most customers.
What are some use cases for HyperConvergence?
Initially one of the strongest use-cases for a hyper-converged infrastructure was to support virtual desktop infrastructure (VDI) workloads; this was due to the need for dedicated storage and compute platforms needed to support most VDI environments. Since hyperconverged architectures generally scale quite linearly, and provide predictable performance, this was a perfect match for those needs. As the hyperconvergence market has matured, in the past 5 or so years many other use cases have been identified and are now widely used such as: disaster recover site equipment, replacement for virtualization “3-tier” infrastructure when both server and storage environments are aging out, ROBO deployments where management of that infrastructure needs to be remote but the remote equipment needs to be resilient, and increasingly for customers looking at Public Cloud, most of the top-tier HCI vendors provide seamless solutions to connect with Amazon AWS, Microsoft Azure, Google Cloud Platform, and others so that predictable workloads can remain on less expensive on premise HCI nodes, and more elastic applications can reside in the cloud, but under the same management pane of glass. In short, with advances in server infrastructure such as SSD or NVMe storage, high core count CPUs, and dense memory options, any workload (including bare metal) can become part of an HCI today.
Who are the leaders in HCI?
Increasing interest over the past few years have shown a number of niche players with more application focused solutions, or some of the traditional IT vendors enter the HCI market with options that may be attractive to your organization due to either existing relationships with a vendor, comfort with their instantiation of an HCI solution, or from some excellent available testimonials from industry/market peers. The current HCI leaders that definitely warrant a closer look include:
- Nutanix Acropolis-based solutions provide end-to-end single pane management and your choice of hypervisors (including their own, AHV), and OEM community hardware platforms
- VMware vSAN/VxRAIL “ready nodes” or appliances for a familiar VMware experience
- HPE SimpliVity – VMware or Hyper-V powered solutions with massive data reduction capabilities
Other industry offerings may be a good fit for an individual organization’s specific needs, however the three mentioned above are regarded as the “thought leaders” and have a well-established ecosystem of support and significant R&D being invested into their current and next generation solutions.
How can Summit Partners help?
Summit Partners believe in working with our prospects and current clients to discover where they are at in their own private cloud journey. We can help at multiple stages in the process, such as evaluating solid options for their particular environments, as not all customer needs are the same. We feel it is always best to have a face to face meeting to walk through your current environments, discuss specific needs, and provide valuable input into your technology acquisition process. We can assist with sizing best fit solutions based on this input, providing highly competitive proposals, and importantly we stand behind the solutions we sell through assistance with site preparation, installation, and knowledge transfer throughout the project. Please contact your Summit Partners account executive to setup your Hyperconvergence readiness review today!