Every year or so the leading analysts of our industries (I’m looking at your Gartner, Forrester and IDC) come up with new (mega)trends that will disrupt our industry. Since I’m a no-nonsense kind of guy (being Dutch will do that to you) it often feels like hot air. They usually start first in the consumer space and later bleed over into the enterprise space, in this case End User Computing – the textbook example of the effects of Consumerization of IT. Smartphones are an often used example but AppStores are a good example too. Consider that the Apple AppStore launched in 2008 and today, slowly, we are seeing AppStores being adopted more and more in Enterprises–7 years later.
This year and next year, however, a lot of these (sub)trends will have a tangible impact on your End User Computing work and in these articles I want to share some of that impact with you.
In this set of articles (in no particular order) we’re going to talk about some of the trends that tangibly impact our industry today or tomorrow for the bulk of our use cases across the world. I specifically say across the world because there are still very different new tech adoption cycles across the world. Typically, the US is the first to adopt new(er) technologies followed by Northern Europe about 1-2 years later. APACJ usually follows 1-2 years after that, followed by the rest of the civilized world (I’ll leave it to yourself to define “civilized”). Also know this article is not about the Workspace of 2020 and beyond where you might talk to Siri/Cortana/Echo/GoogleNow to get your work done and have your lunch in virtual reality, so keep that in mind.
The first trend I want to explore is hyper-convergence and how it will affect our industry in the coming years for the bulk of the use cases around the world.
It’s not about the hardware
I’ll start of by saying that I am not a hardware guy. Trust me, I know. I worked (full disclosure) at a company that is big on hardware, and I can decisively say I am not a hardware guy. Ironically, the fact that I am not a hardware guy actually proves the impact that hyper-convergence has on the End User Computing industry. Let’s face it, the important parts of End User Computing are about software, not about hardware.
That is, except for desktop virtualization.
Pretty soon after VDI started to become popular, many admins would wake up screaming four letters in the middle of the night. You guessed it: IOPS! The storage wars ensued and all of a sudden one had to be a storage expert to be able to host virtual desktops successfully at any scale. That’s just silly. It would be the same if you were expected to be able piece together your own engine if you wanted to drive a car. IT departments want to focus on delivering the perfect workspace with the right applications, data, security, etc., not what the write penalty will be with Office 2016 and RAID 10+11-5. As it turns out, many of the vendors in the hyper-converged space feel the same way.
VDI leads the way
Many of the vendors in the hyper-converged space have a had their early success with desktop virtualization. That makes sense because desktop virtualization, or more specifically VDI, provided unique challenges for the storage that already existed in most datacenters at the time. VDI requires more storage and more IOPS with very high peaks and deep valleys.
Since the good hyper-converged vendors (in my mind) are really about software, they’ve been in a good position to solve the ‘VDI storage problem’ with software-based features like: (inline) deduplication, storage tiering, compression, read caching, and write optimizations. Additionally, these hyper-converged vendors often offer easy ways to get high availability and DR that make the whole persistent vs non-persistent VDI debate a lot less hard from a cost perspective (the management part of it is a separate story).
Another great benefit of hyper-convergence is how easy it is to scale up (often referred to as Web-scale IT). All you need to do is add another unit/node to get X% more capacity while still managing it all from the same interface. According to the hyper-converged vendors this can be done at a lower cost than ‘traditional’ solutions.
If you decide that desktop virtualization is part of your End User Computing strategy, then you should not have to worry about the physical hardware. This is what hyper-convergence in End User Computing is all about to me. Two of my personal favorites are Atlantis Computing and Nutanix.
Atlantis Computing most certainly is a great example of a company that had their early success in VDI and most certainly have the right DNA. I remember back in 2008 when I was at Provision we talked to Atlantis about integrating with ILIO (which actually had layering capabilities as well!). Actually, it is only fairly recent that Atlantis entered the hyper-converged space with their announcement of their HyperScale hardware platformearlier in 2015.
While Nutanix is more of a new player compared to the old-guard of storage (shipping their first hyper-converged product in 2011),Additionally, many people are learning that other workloads in the datacenter can also benefit from hyper-convergence and as such you will see (even) more of it in your datacenters they have seen has seen some amazing growth (they are one of those ‘Unicorns’) and have a valuation of $2B+. What I like about Nutanix are their management capabilities and their drive to make both the datacenter infrastructure and the virtualization infrastructure ‘invisible.’ Of course, keep in mind that these are just two companies of dozens more.
I think we will see hyper-converged hardware having a big impact in End User Computing projects because of the increased mindshare, commoditization, and lower prices due to increasing competition in this space. Additionally, many people are learning that other workloads in the datacenter can also benefit from hyper-convergence, and as such you will see (even) more of it in your datacenters.