Hyperconvergence Explained

Hyperconvergence is But One Option Among Many


[dropcaps type='square' color='#ffffff' background_color='#e04646' border_color='']A[/dropcaps]s you read the various blog posts and articles here at hypercomverged.org, you may assume that the authors feel that hyperconverged infrastructure is the only way to build a modern data center. Nothing could be further from the truth. In fact, I believe that hyperconverged infrastructure is just one of many different options that data center managers have at their disposal. Just like every other technology out there, companies must weigh the pros and cons of hyperconverged infrastructure in an attempt to determine whether or not it’s the right fit for their individual situations.

Hyperconverged Infrastructure Pitfalls

You can learn all about hyperconvergence benefits in the Hyperconverged Infrastructure for Dummies book or by reading other posts at hyperconverged.org. However, education about a particular solution also includes understanding the potential downsides to that solution. Like all things, hyperconvergence has its share of downside that must be considered.

Linear Scalability: A Pro But Also A Con

First of all, for many, hyperconverged infrastructure’s focus on linear resource scalability may be considered a downside. As you may know, when you add additional nodes to a hyperconverged environment, you’re adding additional storage, additional RAM, additional compute, and additional network throughput… even if you just need some extra storage capacity.

This is by design.

It’s no accident that all of these resources are added when you add a node. For hyperconvergence vendors, one of the key benefits that is often touted is this very linear scalability. It’s often expressed as the ability to add resources without running into unexpected resource constraints. For example, as you add disk shelves to an existing SAN, do you run the potential of overwhelming the head unit or the storage communications fabric? There are customers out there that don’t want to have to worry about these kinds of things.

This is the issue that hyperconverged vendors are attempting to solve, but it can come at the cost of wasted resources that you may not need. To mitigate, most vendors in this space now provide differentiated expansion offerings. So, if you need extra storage, buy a storage-heavy node. Regardless, though, you will still be adding additional compute, RAM, and networking at some level. For some companies, this may not be palatable.

‘Ware Outlier Use Cases

Hyperconverged infrastructure is great for most mainstream and even some emerging workloads. However, there are always outlier use cases for which a venture down the hyperconvergence path would be a mistake. For example, if you have an application that absolutely requires dozens of solid state disks to perform well, hyperconvergence is probably not your solution. If you have an application that is seriously imbalanced in terms of storage capacity needs – as in, it requires gobs and gobs of storage – hyperconvergence may not be your solution. It may make more sense to continue to operate with a data center in which you can deftly adjust the size of each individual resource.

Converged Infrastructure

For many, hyperconverged solutions may seem too small to deploy in an enterprise world. Before I continue, I would recommend that you at least look before making that determination, though. However, if you still want some of the procurement and support benefits that come with hyperconvergence but you want more control over the individual resources, look at one of the big block converged infrastructure options on the market. Or, if you want to look at hyperconvergence, but can’t find an pre-built appliance that fits your needs, look at one of the hyperconverged software only options. These take a bring your own hardware approach that may allow you to forge ahead.

DIY Is Still In

At the end of the day, what matters is that your data center environment simply needs to meet the needs of the business. To that end, the traditional do-it-yourself model is still perfectly viable. If it’s working for you, you don’t need to change it just because new architectural options are hitting the market. However, if you do eventually hit resource and performance constraints or you need to simplify what you’re doing, you might consider hyperconverged at that time. Or, you may find a single use case – maybe VDI – for which hyperconverged makes the most sense and you leave the rest of the environment running as is.

Summary

There are certainly a lot of positive reasons to consider hyperconverged infrastructure, but that doesn’t mean that it’s the perfect solution for every company. Do your due diligence and make sure that whatever data center option you choose, you choose the one that best meets the technical, operational, and budgetary needs of your organization.