Systems

Hyperconverged Infrastructure—Summary

Hyperconverged Infrastructure—Summary

Here’s the fifth part of the series I’ve been writing regarding HCI. Again, I don’t want to sway my audience from one platform to another. Rather, I want them to follow the edict caveat emptor. It’s mission-critical for the future of the environment we’re talking about here for the buyer to be aware of the available options. You need to evaluate the functionality you require as well as those covered in the existing products and their roadmaps, and determine which platform best satisfies your needs.

In some cases, the “business as usual” approach may work best—the virtualization platform, whether VMware, Xen Server, KVM, or OpenStack, and even some container-based platforms, must be maintained, expanded, or leveraged for the scalability and approach the business requires. To be clear, most of what can be supported on HCI may be accomplished (within certain parameters) on a more piecemeal virtualization approach. For example, should you choose to use Acropolis as your virtualization platform, you’ll require a Nutanix environment. HyperCore (a KVM-based hypervisor) will point you toward Scale Computing. Again, the goal here is to ensure the alternative hypervisor satisfies the needs. At one point, I had heard a significant percentage of the customers who began with Acropolis subsequently migrated toward VMware. I’m unsure if there are stats on where it stands today, but it’s important to note. Remember, this is no disparagement of Acropolis, but it’s not necessarily right for everyone. The relevancy of this previous statement, of course, is to know your virtualization requirements, and try to ensure you’re covering those bases.

I’ve stressed scalability from both within (storage horsepower versus CPU horsepower), and how these environments can expand from within these categories, or not. It may or may not be relevant for the customer’s needs to choose a platform capable of expanding incrementally in those categories. Should it be relevant, the choice is beholden upon the customer to ensure they’ve chosen the correct platform, because the cost of these choices can be profound.

I’ve also stressed how the storage environment within an HCI platform may or may not have features important to the overall strategy. Are deduplication, compression, and replication part of your requirement set? If so, surely this need will affect your choice. I want to stress how important this requirement can be, particularly to overall backup/recovery/DR needs. In this case, it’s important for the customer to take these issues into account. Knowing your concerns will determine whether this is relevant for your strategy.

Remember, the goal here is to build a platform to satisfy these requirements for the entire depreciation period, so you’ll be able to support your requirements until the equipment reaches end of life.

With no bias toward or against any platform, the true value of how you choose to go with this, or even if you choose not to go with this, will be determined by scalability, management, storage, and ultimately how functional the environment is moving forward.


Matt is a long-time engineer, and presales architect specializing in storage, virtualization, and orchestrations in the cloud. He has spent the last 10 years in the virtualization space focusing on Citrix and VMware solutions based on various storage platforms and specializing in Open Solaris, ZFS, EMC, NetApp, 3Par, HP, and Equalogic.