Image Image Image Image Image Image Image Image Image Image
Scroll to top

Top

2 Comments

Scale Is My Main Concern With Hyperconvergence, But Not In The Way You Might Think

Scale Is My Main Concern With Hyperconvergence, But Not In The Way You Might Think

I’ve worked for a lot of smaller organizations that are often left out when it comes to new trends by companies that are looking for the big enterprise wins. From their perspective, it makes sense; after all, enterprises have larger budgets and can provide vendors with opportunities for bigger deals that don’t take the cycles that lots of smaller deals would require.

However, the fact remains that the benefits that can be had with hyperconverged infrastructure – easier deployment, easier administration, and simplified IT – are the very benefits that can have the biggest impact on smaller IT shops.  However, in many cases, a hyperconverged environment, while it may carry with it lower long-term costs, may remain too expensive for smaller organizations to be able to afford the up front buy that’s necessary to expend.  In other words, it would seem that a hyperconverged environment may not scale down to a point at which it’s accessible to many companies and organizations.

But is that really true?  Let’s take a look at a few ways that companies may be able to take advantage of a hyperconverged infrastructure, even if the pricing appears to be out of reach at first glance.

Consider All the Cost and Complexity Factors

The data center is a complex beast, even for the SMB and small midmarket.  There are a lot of different devices and services that comprise the data center environment, including servers, storage (both rotational and solid state), the hypervisor, backup and recovery tools, and a whole lot more.  When all of these services are considered in total, the cost of a hyperconverged solution may look pretty good.

Here are just some of the costs that should be considered when trying to figure out just how much you’re spending on your data center:

  • Servers.
  • Storage.  But, just as importantly, is your current storage meeting your needs, both from a capacity standpoint and well as a performance standpoint.  It can be expensive to add solid state disks to small environments in a way that is manageable, so a hyperconverged solution may be the right choice.
  • Backup and recovery.  Backup software is a must-have no matter the size of the organization.  In some hyperconverged appliance solutions, backup and recovery software is a part of the deal.
  • Replication.  Many SMBs don’t do disaster recovery because it can get expensive.  A hyperconverged solution usually has replication capabilities built in.
  • Complexity.  Does all of your current data center stuff have an easy administrative experience or, at the very least, do you have staff that can keep up with everything?  If not, streamlining the process with a solution that centralizes administration may make the most sense.
  • The people.  If you had a solution that required less in the way of staffing resources, could you redirect some of those staff efforts to more value-add activities?

Before you decide to eliminate hyperconvergence as a possible solution, make sure you’re really considering the full spectrum of costs and needs and choose a vendor that can help you simplify as much as possible.  SimpliVity, for example, includes everything but the kitchen sink in their hyperconvergence solution.

Move Away From VMware and Hyper-V

Both hypervisors and the management tools around them can be a significant expense to some SMBs.  There are hyperconverged infrastructure vendors out there that work with KVM, which is a free and pretty robust hypervisor.  Scale Computing, for example, provides a hyperconverged appliance based on KVM and they’ve implemented as a part of their solution all of the management and data protection constructs that are necessary for the data center.  Nutanix also supports KVM on their hyperconverged platform.  In addition, Nutanix supports Hyper-V.  For some verticals – education and non-profit, for example – Hyper-V carries with it such a low cost that it’s practically free.

Take a Do-It-Yourself Approach

Perhaps you work in an environment in which you’d like to be able to leverage hyperconvergence for its benefits, but you already have a substantial investment in servers that you need to keep using for a few years.  If the servers are from a mainstream vendor and are reasonably equipped, you might consider using one of the software-only hyperconverged solutions that are available on the market.  Such products are available from companies like Maxta and Atlantis Computing.  These solutions allow you to reuse your existing hardware, but get the benefits that can come with a hyperconverged solution, including easier overall administration and linear resource scalability.

Consider Financing/Leasing Your Data Center Investment

Once you consider all of the costs in your data center, including various pieces of equipment, associated maintenance costs, and ongoing software licensing costs, you may find that the overall cost of a hyperconverged infrastructure makes good financial sense, but you’re still not able to allocate sufficient budget dollars to do the initial up front buy necessary to get to the minimal cluster size needed.

This is where the use of leasing or financing can be beneficial and it’s an approach I highly recommend to IT shops that need to keep a level budget without peaks associated with large capital expenditures.  It’s a great way for smaller IT shops to buy into bigger projects that might otherwise be out of reach.

Comments

  1. Great points! SimpliVity also allows you to start with just 2 nodes in a cluster. And both them and Nutanix offer great entry level models that are typically more than sufficient for SMB customers. Also since Maxta and Atlantis are licensed based on capacity you can add a couple SSDs and some low cost NL-SAS drives to just about any rack server to get started. From my experience if you compare any of those solutions to a shared storage solution that provides a decent ability to scale the hyperconverged model comes out cheaper and provides as-good or better performance.

  2. Agree, great article Scott and thank you Brad for calling out one of that Maxta as very easy to get started with since we run on existing x86 servers. I also wanted to mention that for companies wanting the latest and greatest technology, Maxta is the first hyper-converged solution to run on the new Intel Xeon E5-2600 v3 processors. MaxDeploy Reference Architecture testing shows over 40% improvements over prior gen Xeon processors in an Exchange environment.

    For those concerned about the other type of scale 😉 in the same software release, Maxta was also the first to launch support for Stretch Metro Cluster.

Submit a Comment