ra2 studio - Fotolia
In making end-of-year predictions, of course you want to look back and relive the ones you got right, but we all swing and miss, too.
For example, I forecasted that Hewlett Packard Enterprise's (HPE) Moonshot software-defined servers would take off in 2015, but they were largely an afterthought. I was also optimistic that there would be big news around desktop as a service (DaaS) and application refactoring, but both technologies had a quieter-than-expected year as the major desktop virtualization vendors turned their attention elsewhere.
I'm keeping those last two predictions in play for 2016, but let's take a look at why they didn't work out in 2015.
Microsoft DaaS waits in line behind Windows 10
Microsoft didn't release Windows 10 until late July, so the company had its hands full for more than half the year focusing on a make-or-break operating system. There just wasn't enough time left over to build a DaaS platform and add a Service Provider License Agreement, or SPLA, in the same year. Given the positive response to Windows 10, I'm confident this prediction still makes sense, especially when you consider that Microsoft now allows client versions of Windows on public clouds. That's a step in the right direction and possibly an indicator of more changes to come.
It's hard to understate the potential effect of a Microsoft DaaS platform. Sure, other DaaS providers are fine-tuning their own platforms, but they're unnecessarily complex due to the arcane licensing restrictions Microsoft places on where Windows can run. Microsoft can change those rules and facilitate wider adoption of DaaS any time it wants, but why squander a built-in strategic advantage? If Microsoft changes its licensing rules without having its own DaaS platform ready, it lets competitors build a head of steam.
In 2016, you can expect that Microsoft will release a DaaS platform in just about the same press release as it relaxes its licensing.
Nobody bought an application refactoring company
Application refactoring allows you to convert legacy Windows and Web apps to a mobile interface on the fly using elements of desktop virtualization and without needing access to source code. App refactoring is still a growing market, but a handful of companies stand out: PowWow Mobile, Reddo Mobility, Capriza and StarMobile. A fifth vendor, hopTo, focuses less on converting apps to a native mobile interface and more on creating shortcut bars to make legacy apps easier to use. There was a reasonable chance Citrix or VMware would scoop up one of these vendors in 2015, but as with the Microsoft DaaS prediction, my timing might have been off by a year.
Two app refactoring products in particular -- PowWow and Reddo -- focus on Win32 applications. PowWow reads information from the Remote Desktop Protocol to determine what an app is trying to do. Both vendors' methods have close ties to VDI and Remote Desktop Session Hosts, so it's a surprise Citrix and VMware haven't kicked the tires yet. Citrix has been trying to do something similar to app refactoring for five years -- unsuccessfully -- with its XenApp Mobility Pack software developer's kit (SDK). That technology never took off because the SDK requires access to source code to tweak applications and is limited to only some apps. True application refactoring, on the other hand, can work on any app.
We're really only two years into app refactoring, but perhaps as the products mature and the cream rises to the top we'll see Citrix or VMware absorb one of the existing vendors.
HPE Moonshot runs out of thrust
Moonshot is HPE's attempt to take the "virtual" out of virtual desktop infrastructure. With HPE Moonshot, you can install 45 ProLiant server nodes on each 4.3u chassis. One node contains all the CPU, GPU, memory, storage and networking to power four persistent desktops, and since Moonshot is compatible with XenDesktop and XenApp, it can allow 180 XenDesktop users on one chassis -- all with dedicated resources on a 1:1 basis. Just as intriguing, Moonshot supports between 40 to 50 XenApp users per node, or roughly 2,000 users per chassis.
Sounds great, right? If you're really good at building desktop images, but you don't want to take on the challenging task of sizing storage and building out a virtualization infrastructure for desktops, Moonshot looks to be a fine alternative. So why haven't we seen a lot of growth in terms of number of users?
Hyper-convergence was hardly a blip on the radar when HP developed Moonshot in 2013, but it has since become its biggest roadblock. Hyper-converged infrastructure (HCI) products take away the necessity for you to become a virtualization and storage guru, without making any sacrifices. You still get all the benefits of a scalable, dynamic infrastructure that can deploy on premises, in the cloud or anywhere in between. Compared to HCI, Moonshot suddenly looks old.
It's hard to say what 2016 will bring for HPE Moonshot. I know some huge fans of it, but they're integrators and consultants, not customers.
Even if you're reading the tea leaves closely, the desktop virtualization industry is hard to prognosticate. None of those 2015 predictions panned out, but at least for two of them, there's still a chance in 2016.
Microsoft disses DaaS with Azure RemoteApp
DaaS held back by Windows licensing rules
Refactoring modernizes Windows legacy apps