BACKGROUND IMAGE: iSTOCK/GETTY IMAGES
Hyper-convergence is becoming more ubiquitous in today's data centers. Each vendor positions its product differently. With a multitude of hyper-converged vendors in today's market, this article will help you decide which features and functions are most important to meet your use case and budget.
In examining your current and planned environments, use the answers to the following eight questions to see how systems from varying hyper-converged storage vendors stack up.
How large does your environment have to scale?
Hyper-converged storage vendors have different standards for how many nodes or appliances their products can support in a single cluster and how well they can integrate functions across multiple clusters. For instance, a majority of vendors limit you to a minimum of three nodes. Understanding the minimum and maximum number of nodes will help you right-size your hyper-converged hardware. Similar information for memory, storage and compute capability is also necessary to map expected and projected workloads.
What hypervisors does the hyper-converged system support?
Most hyper-converged storage systems support at least VMware vSphere and Microsoft Hyper-V, while others support more. Ensure the products that make your short list support the hypervisors in your environment.
How does the product integrate with your current and planned infrastructure?
Consider this a variation on the previous question, except here it concerns your existing data center or cloud infrastructure, especially where its management, configuration and security tools are concerned. Your short list should include platforms that will integrate with your current setup easily and support existing technology within your infrastructure. This is important because it will have a huge impact on both the future cost of operation and related personnel costs.
What kind of training, support and maintenance is required?
Hyper-converged storage vendors differ as to the kind and cost of support they make available. Some address support through a separate, renewable contract with a specific term or duration. Others include support in the purchase price -- often for the first year -- with a renewable contract thereafter.
Availability of training, 24/7 support access, consulting services for installation and configuration, troubleshooting support -- including access to user forums, support wikis or knowledge bases and so forth -- are some important factors to consider. Other important factors include cross-shipment of loaner or replacement equipment while appliances are being repaired and escalation of support during outages or planned peak usage periods. If you're going to hang your business on hyper-converged hardware, you must be prepared to see it properly supported, especially when you're under service-level agreements for uptime, response time and so on.
What variations on standard building blocks does the product offer?
Some hyper-converged storage vendors offer only identical modules, with options for memory and processor configurations, but that are otherwise standard for storage and networking. Others offer basic building blocks, but offer storage- and compute-heavy modules -- or even graphics service modules -- for virtual desktop infrastructure. You should consider how your use of compute, storage and networking resources is likely to grow and then decide whether the building blocks that characterize any particular product assemble efficiently to match that growth. Otherwise, you may find yourself tied to a system that isn't flexible enough to accommodate your growth and scale-out without forcing you to waste resources.
What are the availability, disaster recovery and backup capabilities?
Before buying, assess your reliability, availability, failover and disaster recovery requirements and compare them with your users' SLAs and performance profiles. It's absolutely critical that the vendor's offering meet or exceed those requirements without breaking the bank. Researching this with vendors will take considerable time, but that effort is eminently justifiable.
Expert George Crump explains why integration is a common hurdle for first-time hyper-converged appliance users.
It's also a good idea to request access to reference accounts or flagship users and to get the vendor's information confirmed or corrected by a real-world deployment. Some people in your industry may see a competitive risk in cooperating with such information requests, but you should try to find a situation as close to your own as possible in seeking out such a reference.
How easy is it to use and administrate on a day-to-day basis?
It's essential to be able to live with the hyper-converged system long term, so you must understand how difficult it is to install, migrate, manage and maintain. How well does it fit your IT staff resources and their technical backgrounds and abilities? Is training available to help bring staff quickly up to speed? Is a certification program available to ensure that the necessary technical skills and the knowledge needed for administration are documented, transferable and kept up to date? Because hyper-converged products involve technologies that are usually managed by separate teams, who will manage these systems? It might be your server, storage or virtualization team or you may need to reorganize your IT staff to accommodate converged infrastructure. You'll also want to look into the usability of the management environment, what kinds of automation support are available and how accommodating the vendor might be in adding new items to its automation library or run book routines.
How well do costs fit your current budget and future growth plans?
Properly understanding the cost information for a hyper-converged storage system can be a challenge. In addition to the capital expenditure involved in buying -- or recurring operational expenditures if you elect a service-based product -- there will be additional Opex costs for annual maintenance and support. Depending on your requirements, you may also need add-ins to support high availability, failover or fault tolerance, backups and DR.
Learn how the nature of hyper-converged infrastructure has changed
Read more about the progression of hyper-converged architectures
Hyper-converged products move beyond partnerships