In this post, I will look back on and grade my predictions on what would happen with Hyper-V and Microsoft cloud computing in the year of 2017.
Every prognosticator projects themselves as a professional or expert. But how expert are they? (American) Football fans sit through months of predictions about how good college players will be in the NFL after they turn professional. Those same experts forget those predictions as soon as players sign their contract and either defy expectations or stink up the stadium.
Me? I like to have a little fun at year end. This predictions lark thing is more of a laugh than anything someone should plan their budget on. So I’m going to take last year’s predictions, dissect them, and give them a fair score. Let’s see how I did.
I expected that Microsoft was going to push Azure Stack hard at Ignite.
My bet is that at Ignite in September, Microsoft will spend a lot of time talking about its unique offering, which is a common platform (Azure Resource Manager) that can run on Azure in Microsoft’s cloud, or Azure Stack in a customer’s private cloud or a hosting partner’s public cloud.
The opening keynote by Satya Nadella contained almost no product information. A following keynote by Scott Guthrie did contain some information on Azure Stack. As Ignite attendees will tell you, the meat of the conference is in the 75-minute breakout sessions. There were at least 21 such sessions focusing on Azure Stack. Let me give you a comparison. There was just one pure session on Hyper-V, which was about features such as shielded virtual machines.
Meanwhile, over the in the expo hall, DellEMC, HPE, and Lenovo were all making a big deal about Azure Stack. I talked to a few readers and social media followers and the thing a number of them wanted to learn more about at Ignite was Azure Stack.
I think it is fair to say that Azure Stack did get quite a bit of fanfare. Was it a massive “hey everyone, look at this!” … No, it wasn’t but Azure Stack got lots more attention than many other parts of Windows Server, System Center, or Azure. So I’m giving this prediction a solid B-plus.
Nano Server was the headless (or GUI-less) installation option of Windows Server 2016 that Microsoft pushed on us for deploying things like Storage nodes or Hyper-V hosts. Nano Server was also the mini-kernel for containers. I thought that we were going to see a continuation of the “those of you who use GUIs are bad” that a certain Technical Fellow and his PowerShell acolytes have been rehashing since the majority of Windows Server customers reject that line with Server Core.
Microsoft will continue to recommend Nano for Hyper-V hosts and storage nodes and assume that this is what we are deploying as default.
That all changed on June 15th when Microsoft announced changes to Windows Server. A new semi-annual channel would allow Software Assurance customers to opt into bi-annual upgrades. Tucked away in the news was the announcements that the role of Nano Server was to be greatly reduced, along with the installation size of Nano Server. Support for deploying Nano Server for physical and virtual machine roles would be removed from the 1709 and onwards releases and support for the original (single) release of Nano Server (WS2016) would end in the Spring of 2018. From now on, Nano Server was for container deployments only. Microsoft made Nano Server even smaller by removing the physical/virtual roles.
From now on, if you want a Windows-less Hyper-V, you have to continue using Server Core or (the free) Hyper-V Server. The last time I ran a survey on this topic, with a small sample of 424 respondents, over 70 percent of respondents said that a full installation was their choice, no Server Core.
Although I agreed with the many who did not deploy Nano Server, I was wrong with this prediction. I did not expect such a reversal from Microsoft. So, I have to give myself an F.
If you haven’t heard, and most of you haven’t, Storage Spaces is the software-defined storage that has been in Windows Server since WS2012 and improved in WS2012 R2. It then added hyper-converged infrastructure in the form of Storage Spaces Direct (S2D) in WS2016. Storage Spaces should have a huge impact because it reduces storage costs and improves performance. Last year, I said that:
Despite all the promise of Storage Spaces Direct, I don’t see this changing. S2D adoption will grow in cost-sensitive companies that are rich in engineering knowledge, such as service providers, but SAN will continue to rule the roost.
As far as I can tell not much has changed. The smaller vendors in this space (not HPE, Lenovo, or DellEMC), have been the leaders in the market. Meanwhile, the big 3 pay lip service to Storage Spaces at Microsoft events and in announcements. However, on the ground, sales teams continue to be motivated to sell proprietary SANs. Let’s be frank here; Michael Dell didn’t complete the largest tech acquisition in history to sell JBODs and commodity storage instead of EMC storage arrays. And to be even franker, I haven’t heard of a single Microsoft subsidiary that has even promoted Windows Server 2016. Very few bothered with Windows Server 2012 R2 either. If the customer doesn’t know the tech exists, how are they going to be educated to ask about or buy it?
Sadly, this pretty amazing tech still resides in the shadows of Azure, Office 365, VNX, Compellent, and 3PAR, and I deserve an easy A-plus for this prediction.
I predicted that the parts of Azure that would get the most adoption are in the infrastructure space (IaaS):
The leading solutions continue to be new online applications built on IaaS instead of PaaS, and hybrid solutions such as backup and disaster recovery. These are the top three things I see going from conversation into deployments, and I don’t think that will change in 2017.
Microsoft doesn’t even share actual sales figures for Azure, so they sure as heck don’t share information on what parts of Azure are being consumed the most. We can only go on anecdotal and subjective observations.
I stated back in 2016 that Microsoft was struggling to reach out to enterprise developers. I’ve started to reach out to the developer market in my day job, and it’s very clear that Microsoft has little to no relationship with developers outside of the Build conference and Ignite. Developers use Visual Studio and maybe Visual Studio Team Services, but they know little of what Microsoft has to offer. Microsoft had an organization called DX to work with that market, but DX was an abject failure as far as I saw, focusing on tiny verticals and horizontals that had little potential. Worryingly, for Microsoft, the default cloud position for developers is AWS, despite the huge efforts on making coding easy, serverless computing, DevOps, open source adoption, and more that Microsoft has done in Azure.
From what I see in my market, the activities in Azure are all built mainly on Infrastructure – virtual machines. In fact, there have been shortages in the European regions of Azure for some mid-level host types, such as the D_v2- and the F-Series families. I’ve had customers that couldn’t deploy machines in these series and I’ve read headlines in the news about these shortages. Emails from Microsoft blamed unexpected demand.
I think we will see how mid-large companies will embrace cloud transformation in 2018 and understand that it’s more than “lift and shift”. Until then, I can only give myself an A on this prediction.
For a while there, I think most of us were waiting to see a Surface Phone. And to this day, there are members of the media that still cannot face the truth. I predicted that:
The hardware will look amazing. The spec will be out of this world. … And the sales of Surface Phone will suck.
So, there was no Surface Phone and that will dent my grade! In fact, Microsoft pretty much killed off Windows Mobile/Phone or whatever it was last called, so I was at least right about the “fail” bit of the prediction! I’ll go with a D because somewhere in a secure facility in Redmond, I’m sure there’s some Surface Phone hardware that will never see the light of day.
Petri probably could have hired a chimpanzee to make these predictions and maybe it would have done just as well. I think I will give myself a C-plus and that might be generous.
If you’d like to see more predictions on what I think is coming in 2018, then watch out for another post before the end of the year.