There were dozens of announcements at Microsoft’s Build conference last week, but perhaps one caused the most angst among the cloud cognoscenti.
I'm referring to the upcoming general availability of Azure Stack. Microsoft’s offering will let organizations leverage the Azure cloud operating system, but only within the context of an on-premises deployment.
Azure Stack has something of a checkered past -- it has been announced, in one guise or another, more than once. I remember years ago the notion of a private cloud deployment that would involve Microsoft software and partner hardware. That never really eventuated, and things went quiet.
Then Azure Stack was resurrected and the release date was set (and then changed). The announcement at Build that Azure Stack is in the final technical preview stage indicates that the release will be, as indicated, the middle of this year.
This is where things get interesting. One Twitter account, who I’ll not name but who, suffice it to say, has a habit of anonymously popping up and crucifying vendors he or she doesn’t agree with, piped up suggesting that Azure Stack is the product that will completely decimate Microsoft’s cloudy ambitions. That’s a pretty impressive statement given that pretty much everyone agrees Microsoft is the second largest public cloud vendor on earth, perennially bridesmaid to Amazon Web Services. To think a single product could knock it off that perch seems a little but like hyperbole to me.
The said critic suggested Azure Stack is essentially the same thing as Oracle’s “cloud in a box” offering and, as such, does nothing but give those “server huggers,” the IT professionals who feel threatened by the cloud and want to stop its progress as much as possible, an out. Essentially the perspective was Azure Stack is a false cloud and will allow conservative organizations to avoid moving to the real cloud.
Sigh.
Let's get this straight: Azure Stack is a specific offering that is targeted for specific use cases. If you are, for example, running a mining operation deep in the Congo, you’re potentially going to have some issues -- both in terms of latency and outright connectivity -- leveraging the public cloud. Even more relevant, if you’re running a shipping company, the public cloud is an awesome thing, but trailing a fiber optic cable behind your ship as you cross the Atlantic, simply to retain connectivity, doesn’t really work.
And, despite the howls of derision from those who opine about these things, there are some use cases where either sensitivity, regulation or some other factor means that siting data behind the firewall is preferable.
Now, I can just hear the criticisms from some quarters who will suggest I am a private cloud apologist and not really committed to the value the public cloud brings.
My response? I’ve been evangelizing the cloud pretty much since before the term was invented. I fully believe the public cloud should be the default way for consuming technology -- whether it is infrastructure, platforms or software. But, I’m also a pragmatist and can accept that sometimes, for a myriad of reasons, that won’t work. That’s not to say that I’ll shy away from encouraging recalcitrant organizations to cease denying the existence or validity of the cloud. But where an organization can show a valid reason to not go public for a particular situation or use-case, I’ll back them.
A good case in point is one introduced by Microsoft at the show. To my point about shipping companies, Carnival Cruise Line is deploying Azure Stack on some of its ships to allow the organization to write an application once and then run it either in the cloud or on-premises while on the water.
There is an interesting habit within the technology industry -- one of denying the validity of an operating model or approach that doesn’t meet our tightly constrained ideas of “best practice.” This attitude is both unrealistic and fundamentally disrespectful to the end users we all purport to be building solutions for. The criticisms around Azure Stack are an example of this, in my view, and need to go away. Azure Stack isn’t for everyone -- indeed the vast majority of Azure customers have no need for it -- but, for those who do, it’s going to be a very welcome product addition.