We are privileged see a lot of very cool, ground breaking, technology every day here at RedMonk. Occasionally though, we go to a technology event and have a genuine wow, that is incredibly cool moment. One where you see a piece of technology that has the potential to have a really positive impact on peoples lives.
I had one of these moments at Microsoft Build 2016. If you only do one thing from this post, watch the video about the work Microsoft engineer Saqib Shaikh is doing:
Yes, this is marketing, and the video came at the very end of Satya Nadella’s keynote. However, marketing aside, there is something wonderful about seeing a number of technologies converging into something that can make such a significant difference to peoples lives.
Bash on Windows
To say that using the phrase “bash on windows” underplays what Microsoft and Canonical are working on is, to put it politely, a significant understatement. Essentially anything packaged for Ubuntu that does not require a kernel hook is now available on Windows. Let us be very clear, as a number of people reacted on twitter during the event stating we have had bash on windows for years, this is much, much more than what was there before.
The implications of having apt-get available natively on Windows are huge in terms of developer stickiness. One of the problems Microsoft have had over the last few years is convincing developers to return to the Windows platform, and ensuring those that are not constrained by corporate standards remain. Think of people such as node.js developers, lamp developers and so forth.
When you see people like Laurie Voss, CTO of NPM tweeting things such as:
If Windows had had bash support 5 years ago I probably wouldn’t have left. It is a pretty stunning and laudable turnaround for Microsoft.
— Laurie Voss (@seldo) March 31, 2016
you know that there is something significant with this announcement.
At RedMonk we are continuously reminding people that packaging wins, and this is a huge step forward for Microsoft in terms of packaging.
One of the next big waves in cloud computing is going to be serverless, and as predicted earlier this year Microsoft announced their offering in this space, Azure Functions. At a very high level serverless computing provides the ability to execute code at specific times (think event driven, e.g. dealing with some sensor data or similar for example) and only pay for precisely the computing resources you are using. This is quite a paradigm shift for a lot of people, but we are already seeing a lot of interest in the overall serverless model.
Microsoft are differentiating their offering from those of Amazon, Google and IBM in a couple of ways. Firstly, the number of languages supported is currently significantly higher than any of the other offerings. node.js is the first language highlighted, as with all the other serverless offerings, but there is also support for C#, PHP, F# and Bash among others. We do expect this gap to be closed relatively quickly if significant traction is seen for languages beyond node.js. Secondly, like IBM, Microsoft have released an opensource sdk, under an MIT license, as well.
On a different note, if you are interested in learning more about the whole serverless phenomenon, check out ServerlessConf in New York next month. In the interests of full disclosure you should be aware that RedMonk is involved in organising ServerlessConf.
Azure IOT & Partnerships
We had the opportunity to spend some time with Sam George who runs the Microsoft IoT team. As we have come to expect from previous conversations with his team, Sam was very clear that their focus in on B2B, with a strong emphasis on developing and enhancing both existing and new partnerships.
Unlike many other companies, however, Microsoft are very interested in the last mile. Given their familiarity with hardware support and device drives over the years this makes sense, and to this end they announced the Azure IoT Starter Kits.
We also got a preview of the IoT Gateway SDK and a further update on their device management capabilities, which we got some insight into from Juan Perez at ThingMonk last year. Both of these capabilities are key to the large scale IoT deployments, and it will be interesting to see their adoption curve over the coming year.
Where the entire IoT story comes together is the integration between Azure Functions, analytics, machine learning – in particular predictive analytics and the IoT hub. Microsoft is far from the only game in town here, many other vendors are operating in this space, but their story is particularly complete in comparison to most others at this stage of the game.
Machine Learning & Cognitive APIs
I have to be honest, like my colleague James Governor, I did not anticipate cognitive sticking as a name for deep learning and neural networks, but stick it has. I still find myself spending an inordinate amount of time explaining what “cognitive” means to people when it comes up in conversations, but perhaps with all of the marketing dollars being spent around the word “cognitive” Microsoft and IBM will make that part of my job easier.
Now stepping away from my thoughts on the word, lets look at the offering which was announced at Build. Microsoft announced a set of apis covering vision, speech, language, knowledge and search. The breadth of coverage is quite impressive, and as I mentioned previously is something I will return to in a future post.
The big negative on the Azure machine learning offering is the lack of availability in many Azure regions. As of today there is availability in three of twenty regions. We do, however, expect this to be addressed relatively quickly, and given where most enterprises are on their machine learning journey, the geographic coverage should be addressed by the time the market needs it.
Bots & Conversation as a Service
The race to take ownership of the bot market has started in earnest. From an over hyped marketing perspective you could say bots are this years IoT. But as a potential inflection point for how businesses and consumers will interact with technology in the future there is no doubt that bots are incredibly interesting.
More importantly you need to once again step back and frame how bots can be used through the wider lens of mobile and cloud, and understanding the importance of workloads. Microsoft are still not, even with the Xamarin acquisition, a significant player in mobile. While this has the potential to change, they have a tough journey ahead.
However, Microsoft provide endpoints for bots to interact with, and they can then integrate these endpoints all the way through to their various machine learning offerings, analytics and so forth. To fully leverage all the pieces, you will first want your workloads and data in Azure, and build from there.
Microsoft are showing a lot of momentum in the right direction, and unveiled some very interesting new offerings during Build.
The focus on working with developer communities, and the overall developer focus that Microsoft have traditionally had is being extended further and further into the opensource world. As I have mentioned in the past there are some significant cultural challenges ahead, and a huge amount of resistance from some quarters. But overall Microsoft are heading in the right direction.
On the business side, conversations we have with C-level execs are far more likely to be positive than negative. Microsoft are providing a path to cloud for their traditional install base, and working very hard to ensure the cloud natives can come along for the ride. Amazon may be the gorilla in the corner of cloud, but in terms of overall enterprise footprint Microsoft have both the presence and resources to directly challenge Amazon.
The competition between the major players will be good for developers and businesses.
Disclosures: Microsoft paid my T&E. Amazon and IBM are current RedMonk clients.