You have to hand it to the Rolling Stones. They foretold the current debate over…
Hyper Converged Infrastructure (HCI) is a type of software defined data center or cloud built on standardized, prepackaged hardware. In hyper converged solutions, each node is identical, with a set combination of compute and storage. A virtualized network is run on top of all the hardware.
Because all of the hardware is standardized, hyper converged infrastructure is relatively inexpensive, and easy to setup, run and service. However, it is not necessarily suitable for large networks or demanding enterprise cloud workloads.
Expert Insight: Randy Downey, Senior Cloud Architect
For many SMBs, hyper converged infrastructure makes a lot of sense. HCIs typically use 8-10 nodes which can support applications like remote desktop environments, development and testing. They run on affordable commodity hardware, which is easy to swap out when it goes bad. However, large landscapes, and high-performance or non-traditional workloads like the SAP HANA cloud, require other data center configurations, such as converged infrastructure.
What is Hyper Converged Infrastructure vs. Converged Infrastructure?
The biggest difference between converged infrastructure and hyper converged infrastructure is in the hardware. In both models, the provider uses a network of machines to create a pool of resources, which can be used to host whatever applications the tenant wants to run. However, in converged infrastructure, the provider purchases servers and storage arrays separately, while in hyper converged solutions, each chassis comes with a standard amount of compute. Converged infrastructure can also be configured and run in ways HCI can’t.
Hyper Converged Infrastructure Limitations
Because converged hosts purchase and install processing power and compute separately, they can optimize their networks for, say, big data storage or processing-intensive applications, or in-memory database systems like SAP HANA that require a lot of RAM. You can’t do this with hyper converged infrastructure, because you have to upgrade RAM, processing and storage together.
Hyper converged infrastructure is also less flexible than converged landscapes. In a converged node with eight servers, you could virtualize five and install and run operating systems on the other three, whereas on HCI, all the servers would be under a hypervisor, which would allocate resources into virtual clouds. This makes converged infrastructure more useful for many enterprise applications with exacting performance, security or hardware configuration requirements.
Hyper Converged Infrastructure, the Managed Cloud and the Economy of Scale
Unless you’re hosting your own onsite data center, hyper converged infrastructure is probably not the best solution at this point, simply because it’s not the most cost-effective way for a provider to provision compute. Both enterprise data centers and commodity cloud providers can run more cost effectively by allocating compute out of large pools of shared resources than by repeatedly building small clusters of nodes.
What is Hyper Converged Infrastructure Going to Become?
HCI isn’t flawed or bad, it’s just new. Hyper converged solutions are already useful for a number of applications (we use HCI to administer our own next generation cloud). As the cost of compute continues to decline and the software used to run hyper converged infrastructure continues to improve, companies will be able to overcome performance, size and configurability limits, while benefitting from the simplicity and robustness of the physical architecture.
Symmetry is committed to staying on the cutting edge of cloud hosting and enterprise application management. Our next generation converged cloud is providing easier provisioning, more flexibility and better ROI, allowing our customers to move from multiple clouds to a single vendor model. When the hyper converged cloud is ready for you, you can bet we’ll be leading the pack.
Contact us to learn what Symmetry’s cloud can do for your business.