Finding consistency within Microsoft's ever-increasing product range

17.08.2015
Microsoft is a company that I've long watched from afar, but never really engaged with much. I've never been able to attend any of its large international events and hence haven't had a really good handle of what the breadth of its product portfolio looks like, especially when compared to other vendors. Recently I was invited by Microsoft to meet with a bunch of its executives to get a deep dive into what it's doing. (Disclosure: Microsoft covered my travel and expenses to visit Redmond.)

It was interesting to visit Redmond at this time since Microsoft is finally being seen in the marketplace as a very real competitor to Amazon Web Services in the public cloud space. I wanted to take the opportunity to hear Microsoft's perspective on what it does and, more importantly, what it does differently from the other vendors.

In an introductory session, Jason Zander, Microsoft's corporate vice president for Azure, explained that Microsoft's perspective on what enterprises want consists of a Venn diagram with three distinct requirements — hyperscale, enterprise-grade and hybrid. Diving into this a little bit, we can see that all of these requirements are steeped in the reality of enterprise needs:

If one feels comfortable about these three requirements for a vendor, it is interesting to think of this in the context of the largest technology vendors. Google, for example, clearly has hyperscale. Arguably it is also getting toward an enterprise-grade offering as it matures in its understanding of what it really takes to deliver enterprise products and services. But it doesn't really play a hybrid story. AWS, for its part, is similar, and while the biggest customers like the CIA can obtain their own private cloud, the average enterprise is getting a principally public cloud story from AWS. True, it enables such things as direct connections between existing workloads on third-party infrastructure and its own public cloud, but the Microsoft view of the world is that enterprises like the idea of having one throat to choke. Looking at the sweep of other vendors in the space, IBM certainly has an enterprise-grade and hybrid story, but it still doesn't have a real hyperscale offering. True, it is building out its SoftLayer public cloud, but it simply can't compete with the scale that comes from running the myriad of massive webscale tools that Google, Amazon and Microsoft do.

A good example of where the hybrid story, as delivered by one vendor, comes in came from a technology that Zander told me about that I wasn't aware of. Using SQL Server Stretch Database, a single database can "stretch" between cloud and on-premises. Thinking about this within the context of, for example, an enterprise with a particular regulatory requirement, this means that an operational database can sit in the cloud, with parts of it (those parts with sensitive data) residing on-premises. It's within this kind of deep-use case that a single vendor offering a consistent hybrid story comes to the fore.

Of course, it's not all roses for Microsoft. The company was both slow to the cloud game and had a few stumbles en route. A few years ago, it announced a partner-led "cloud in a box" strategy whereby partners would be able to run Azure on their own data center resources. While it sounded good, this never really found traction, and Microsoft hasn't yet cracked the nut on the truly globally federated cloud. While it's easy, for example, to say that Australia and New Zealand are both well served by an Azure data center in Australia, in some cases that isn't enough. The partner channel is an opportunity to resolve that, or alternatively some kind of Microsoft-led version of dropping shipping containers within locations that are fully self-contained mini data centers. 

While Microsoft's offerings in this space (its own Azure services, the Azure Stack that partners can download and run on their own hardware and the Cloud Platform System that offers a hyper-converged cloud offering) all sound good, it still feels slightly discontinuous. Yes, it is excellent that the technology running on all three approaches is largely the same, but there isn't yet sufficient quantities of all three approaches in-market to make it really feel mature and consistent.

And in some ways this gets us to the nub of the problem I have with Microsoft generally and Azure specifically. Microsoft has all the component pieces — machine learning, API services, mobile development, databases. At almost every turn, it has a distinct piece of technology that customers want. But it seems to me that those offerings are just that: distinct pieces of technology delivered by an engineering-led organization that is yet to really crack the code on how to tell a consistent story that goes beyond technology, but rather delivers a business message underpinned by a consistent and coherent platform.

It wasn't until my last meeting of the day that I saw this sort of story, albeit in a roundabout way, in action. I had a tour of Microsoft's digital crimes unit, a futuristic location where Microsoft analysts, along with industry partners, do work to track cybercrime and help bring perpetrators to justice. The cybercrime unit tracks millions of license activations, external data sources and applied machine learning, analytics and high-power visualizations over it to develop learnings that, in this case, can be applied to cybercrime problems. Taking a deep dive into the technology stack that powers the DCU, it was an amazing example of a bunch of technologies (not all of them Microsoft's, I must say) combined to deliver a business outcome. It was, at its essence, an opportunity for Microsoft to tell a really compelling story, at a business level, about how its massive portfolio of technology pieces can be combined to deliver business value. The DCU is built upon Azure, SQL Server, the Microsoft Analytics Platform, Power BI, Office 365 and Surface Hub — all integrated to deliver real-world outcomes.

And this is, in my view, the nut that Microsoft needs to crack. It's got the technology pieces, has the scale to innovate with the best, and has the existing relationships it can leverage to be a massively important player in the cloud space. It just needs to learn to contextualize all of that in a way that businesses, as opposed to technologists within businesses, can understand. It is a good problem to have, far better to have the product and lack the good stories than to be full of stories without the product underpinnings. Either way, I'm looking forward to seeing Microsoft shake things up a bit, and balance its long-held technology story, with a newly emergent business one.

(www.computerworld.com)

Ben Kepes