I’ve broken this post into two, to make it easier to digest. Today we’re all about what Azure Stack isn’t, some advice for partners nervous about the future and a look to the future for Azure Stack.

I’ve been excited about Microsoft’s plans to offer an on-premises version of its cloud infrastructure offering for the longest time. While there are many cases of organizations that are completely in the public cloud, for every one of those there are literally hundreds of organizations whose needs are subtly different – organizations with a geographic, connectivity or regulatory reason to retain at least some of its infrastructure footprint on premises. And so the news, all those years ago, that Microsoft would offer some kind of on-premises cloud was both fascinating and exciting.

It’s fair to say that the process hasn’t exactly been plain sailing, and Microsofhasve had a few goes at getting this right. As I’ve written before, I’m also a little worried that there isn’t more clarity about exactly what Azure Stack isn’t.

It is for this reason that I took the time to engage in a lengthy (apologies in advance!) question and answer sessions with Natalia Mackevicius, the Director of Program Management for Azure Stack. I wanted to get deep into the thinking behind Azure Stack, the direction they’re taking it, and exactly what it is (and, perhaps more importantly, isn’t.)

And so, without any further ado, here goes my epic Q&A.

I wrote a post challenging Microsoft to be more emphatic about what Azure Stack isn’t. I perceive that there are lots of people thinking Azure Stack will be the answer to all their problems when, in some cases at least, it won’t. Can you help us all to understand what Azure Stack isn’t – and why that’s important?

When NIST first introduced us to the categories of public, private and hybrid cloud, that was a major step forward in evolving the utility model of computing, but the industry has come a long way since then.

For some customers, the expectation arose that they would be able to create and sustain “alternate” clouds to the public providers. There are very few who have been successful creating a general purpose private cloud. More importantly, most attempts at private cloud result in the realization that sustaining the rate of innovation required to keep that private cloud available, useful, and relevant is a daunting endeavor.

Azure Stack is not an alternative to Azure. Rather, it’s an extension — a way to leverage a company’s existing assets, such as oil rigs, factories or shipping vessels, and integrate them with the capabilities and ecosystem of a major public cloud provider.

There were also some who expected that private cloud would improve operational efficiencies in the same way that virtualization had for the client-server generation of applications. Azure Stack is not an incremental improvement to virtualization.

Most customers don’t need another way to deploy and manage the last thirty years of solutions. In fact, many are actively auditing their IT portfolios to determine how to get out of running legacy systems, so that they can focus on building business solutions based on cloud tools and techniques for the next thirty years.

As they go forward, and invest in cloud native applications, those can run on Azure Stack as needed.

And for those partners who see it as a way to avoid disruption in a world increasingly aligned with the public cloud, what advice do you have for them?

The disruption of cloud is in large part due to the way that it unlocks truly transformative innovation for customers; it fundamentally changes the rules of the last thirty years of IT that business practices are built around. In fact, almost every Azure Stack partner we work with is rethinking their own business models to account for the new rules, such as traditional hosters who are transitioning to a managed services model that’s easier to tie to usage style computing.

I think we’ve come a long way by helping them understand that Azure Stack is an extension of Azure that enables partners to open new lines of business for their Azure practices across the whole of their customers’ business assets.

Many partners understand that they have a massive business opportunity to help customers focus their IT resources on delivering business value through innovation, rather than simply managing costs and risks for heritage systems.

Awesome. With that out of the way, and with a deeper understanding about what Azure Stack isn’t, can you spend some time talking about the benefits you can deliver when Azure Stack is used in the ways it was intended?

For many established companies, there’s a tremendous amount of technical debt that has been built up and is holding back application innovation.

With a consistent platform across Azure and Azure Stack, companies can get beyond their debt by investing in new people skills (for both development and operations), modern application development architectures and processes, as well as operational standards that work the same way wherever they need them.

It gives them a way to not have to split up their IT investments into the different technology silos every time a business need or opportunity arises.

Crystal ball time. We’re sitting here in three or five years. What things do you think Azure Stack will be used for and how will the most progressive companies use it?

I have heard Satya say that computing is going to get more distributed and I agree with that. Azure Stack is a major component of that process.

Customers are going to initially build a solution for Azure and then deploy the exact same thing to Azure Stack or vice-versa.

But soon after that, and we are already seeing customers thinking this way, we’ll see single applications distributed across clouds the same way that many applications are distributed across servers today.

A potential example could be that a customer has a single server or device deployed to many locations, each transmitting IoT data to a regional hub where Azure Stack is running. Azure Stack is used to preprocess the data, meeting some kind of regulatory requirement, and then transmits the transformed data to Azure, where the data from each region is used to train a Machine Learning model. That model can then be brought back down and used to score data in the regional Azure Stack and power a line of business application.

As this model evolves, I think we’ll see new patterns, practices and architectures develop for both applications and data that push the hybrid cloud definition even further.

Ben Kepes

Ben Kepes is a technology evangelist, an investor, a commentator and a business adviser. Ben covers the convergence of technology, mobile, ubiquity and agility, all enabled by the Cloud. His areas of interest extend to enterprise software, software integration, financial/accounting software, platforms and infrastructure as well as articulating technology simply for everyday users.

1 Comment

Leave a Reply