Categories: Analyst Blogs
Tags: blog, converged, Eric Slack, hyperconverged,
First came the Converged Infrastructure (CI), providing a way to simplify the integration of open systems components in large data centers. Then came Hyperconverged, which sounds like an even better solution for these same use cases, but really isn’t. Where CIs combine existing components, Hyperconverged Infrastructures (HCIs) represent a completely new product. These two technologies share the same root word but not much else, and they don’t compete for the same IT projects.
CIs are rooted in the enterprise data center, while HCIs are most often found at the edge, in remote offices, special projects or departments – or in the SMB space. But that may be changing as enterprise IT organizations start to warm up to the Hyperconverged concept. This new technology won’t replace Converged Infrastructures but does it have a place in the enterprise data center?
What is Converged Infrastructure?
Converged Infrastructures (CIs) are bundles of standard storage, compute and networking components that are used to create large data center infrastructures. This architecture combines or “converges” off-the-shelf components at the rack level, typically with a virtualization, management or orchestration software layer, testing and certifying the entire stack to eliminate complex integration. CIs are typically sold as a package with a single product SKU for easy procurement and comprehensive support or offered as a reference architecture that customers or integrators use to assemble a complete compute environment.
One example of a Converged Infrastructure solution is EMC’s VCE Vblock systems which include EMC storage, Cisco servers and switches and a VMware hypervisor sold as a single product that’s supported by VCE. Another is NetApp’s FlexPod, a reference architecture comprised of NetApp storage, Cisco servers and switches and a VMware hypervisor.
The big drivers behind Converged Infrastructures are simplicity of design, ease of implementation and management. Essentially, companies were tired of dealing with the complex integration and scaling associated with traditional SAN-based open systems infrastructures that were common in enterprise data centers. As server virtualization simplified their compute environments they wanted to do the same with their storage and networking.
What is Hyperconverged?
Hyperconverged Infrastructures take convergence a deeper level by combining these same functions – compute, storage and networking – into a single chassis, typically based on industry-standard x86 server hardware. Hyperconverged systems leverage software-defined storage technologies to create a cluster of these modules or “nodes”, each contributing local storage resources to a common pool. Each node also runs a hypervisor instance, used to support VMs and usually the hyperconverged software as well.
While they can be sold as software that’s assembled with the appropriate server hardware by the customer or an integrator, most HCI solutions are sold as appliances by a single vendor that assumes end-to-end support. Examples of Hyperconverged Appliances (HCAs) are Nutanix, SimpliVity and HPE’s Hyper Converged 250. The biggest single use case for HCAs is probably in remote locations (ROBO), but other than that, these products appeal mostly to small and mid-sized companies (SMB) where they’re being used to replace traditional IT infrastructure during the refresh cycle or to support a new application (like VDI).
Hyperconverged vs Converged
While they do share the same root word and a similar primary benefit (reducing complexity) Converged and Hyperconverged infrastructures have very little in common as products or the use cases they support. CIs use existing rack-level IT components in standard configurations to simplify design and implementation in enterprise IT departments. And deployments are decidedly enterprise too, often in environments that are currently using these same products. Converged Infrastructures allow for incremental growth as well, but those increments are typically at the rack-level.
Where CIs are largely a bundling exercise, HCAs are an entirely new product, one that offers a low-cost, easy to implement solution, even by non-IT personnel. They also scale in smaller increments than CIs, at the server vs the rack level. As such HCAs offer a solution that appeals in different ways to a much broader range of companies, starting with the SMB and extending into the enterprise, but for different use cases than Converged Infrastructures.
Hyperconverged Appliances in the Enterprise
As I mentioned earlier, enterprise IT isn’t considering Hyperconverged next to Converged solutions for their tier-one environments. That said, they are buying HCAs for other uses, most notably remote locations where they don’t need much capacity and don’t have IT expertise. Appliance-based solutions are ideal for these ROBO use cases, but can also give enterprise IT a way to standardize their infrastructures that exist outside the primary data center.
All large IT organizations need to stand up projects that aren’t tied to the primary infrastructure. These can be systems that need a different set of technological characteristics, such as VDI, gear that’s not funded by IT, such as customer-owned systems or for projects with security or data residency restrictions.
In these situations a standard solution with a wide range of advanced features that can be quickly configured (and reconfigured) to meet specific requirements is very appealing. If that solution is easy to use and can be implemented by junior IT personnel, so much the better. These are some of the reasons enterprises are looking at Hyperconverged Appliances.
Many products have long lists of features that sound the same but work very differently. It’s important to think outside of the checkbox of similar-sounding features and understand how technologies and products differ.