The fast-growing appeal of hyperconvergence software makes a lot of sense. Compared to the appliance model, which delivers a box with hardware and software bundled together, a true software model gives customers much greater flexibility and freedom. With a software-based approach, customers can pick and choose the best hardware for their needs and incrementally scale and improve their infrastructure without paying for a major upgrade package. At the same time, software vendors have a business model with much better margins than companies selling appliances.

Despite the benefits of the software model, most companies begin with a focus on shipping appliance-based hyperconvergence solutions. It’s simpler to sell, whether directly by the vendor or indirectly through a reseller or distributor, and it’s easier to develop and support the software on a limited number of hardware platforms. These vendors may try to pivot to a software model in order capture the appeal of software’s flexibility — and to improve their profitability — but making the transition from selling appliances is complicated and takes time to develop. More often, vendors will claim to offer hyperconvergence software, but still significantly restrict how their solution can be deployed and used, essentially delivering an appliance in software clothing.

This trend has created a lot of confusion around hyperconvergence software, and what it can and cannot do. Myths about the drawbacks and limits of hyperconvergence software have spread due to the fact that many of these “appliance solutions in software clothing” don’t have the flexibility of a true software solution. To understand how hyperconvergence software really works, and how to spot a “fake” software solution, it’s important to debunk the following myths.

You have to replace your hardware to support new hyperconvergence software.

If a vendor tells you that you need to buy new hardware to put hyperconvergence “software” in production, then they’re not really offering software. While you need to ensure that your existing hardware has adequate specifications to support the software, you shouldn’t have to buy new hardware to switch to a software-defined hyperconvergence solution. Sometimes vendors will try a bait and switch by offering “trial” software that you can run on your own hardware, then force you to purchase new hardware to put it in production. The truth is, you don’t have to buy new hardware to run hyperconvergence solutions from real software vendors.

You have to pick your server hardware from an approved list of server SKUs.

If you decide to refresh your hardware when you implement hyperconvergence, sometimes you’ll hear that you can only pick from a limited set of server SKUs. However, that’s only true if you’re dealing with vendor selling what looks like a software solution, but that is really just an appliance with hardware and software priced out separately.

You can’t add capacity within an existing server.

Appliance vendors want you to think that the only way to increase capacity for hyperconverged infrastructure is to add another appliance. But true hyperconvergence software enables you to increase capacity within an existing server by adding drives in open slots, swapping in higher-capacity drives, or by adding new servers. While some vendors offer less expensive storage-only nodes, the step-up cost of another box makes this more costly than it seems. Only fake software providers force you to add nodes in order to increase capacity.

You can’t buy a lifetime license for hyperconvergence software.

Offering a lifetime license for hyperconvergence software undermines the business model of appliance vendors, so they’ll tell you it’s impossible. An appliance model ties the software license to the physical appliance, so that when you refresh your hardware you’re forced to buy a new software license. This type of “term license” allows vendors to charge you for the same software solution again and again, but it’s not the only option.

Although many software companies now offer term licenses to provide subscription-like pricing, nearly all software companies — including hyperconvergence providers — still give you the option to buy a perpetual license that you own forever. And if you like the idea of owning the software for life but don’t want to pay for it all upfront, you can lease the software from any number of leasing companies to get the best of both worlds.

Hyperconvergence software doesn’t allow you to scale compute and storage separately.

In the same way that true hyperconvergence software allows you to add capacity without adding new nodes, you should be able to scale compute power independently, whether it’s within an existing server or through the addition of a compute-only server. The myth that running hyperconverged infrastructure means that you can’t scale compute and storage independently comes from an appliance approach that always bundles performance and capacity upgrades in a new node.

As the hyperconvergence market moves to embrace a software approach, some vendors will try to make the line between appliance-based solutions and software even fuzzier. Companies that started off selling hardware will try to grab on to the software trend by putting out solutions that don’t look like appliances but still come with many of the same restrictions. As you evaluate hyperconvergence, it’s important to understand what these limitations and related costs will be in four or five years when you need to refresh or upgrade.

The rapid evolution of infrastructure platforms means that the ability to scale, choose, and change hardware — and use different hypervisors — has become more important than ever. If you get stuck with inflexible architecture it will create much bigger problems for IT down the road. Clearing up the myths and confusion around hyperconvergence software will be crucial for you to navigate the rapidly shifting infrastructure landscape.