Microsoft's vision of a data culture is a sign of the times. Drive into San Francisco and there are billboards everywhere advertising tools for working with big data, telling us "We're all data nerds now". That's true -- as the adage goes "you can't manage something until you can measure it," and with today's tsunami of data there's little we can't measure.
New CEO Satya Nadella's vision of a world of ubiquitous computing and ambient intelligence is a way of looking at both sides of the story, where smart endpoints deliver data for analysis, and then use the results of that analysis to help users make decisions. It's a vision that's aiming to bring together the disparate elements of the Internet of Things, with the intent of giving them a purpose. There's no point in instrumenting every tanker car in an oil train if you're unable to use that data to handle routings and to plan maintenance schedules.
But you need tools to both capture and analyze that data -- tools that are able to scale.
That's where Microsoft's Azure cloud platform comes in, as a place to both capture and analyze that data. A new Azure service, the Azure Intelligent Systems Service (AISS), forms the backbone of Microsoft's Internet of Things. Intended to allow organizations to quickly build and deploy sensors and controllers, and to work with the data they produce, AISS builds on Azure's service bus model to provide a way of routing data, and of applying rules to the input and output of data analysis services.
While Microsoft would obviously prefer you to use it with its own Windows Embedded OS, AISS is endpoint agnostic: As long as a device or a gateway can connect over IP to the Internet, it can be connected to AISS. There's no need for new infrastructure, as you're using your existing Internet connection and taking advantage of a scalable cloud service to manage devices and to process the information they deliver.
The Internet of Things needs to be part of a data platform; it's about taking data from not-particularly smart devices, processing it at scale, applying intelligence (whether human or artificial), and then acting on the insights received. There's a lot that needs to be done to build such a system, and Microsoft is betting that its existing tools and services will be easier to use than building your own. Certainly it's an approach that should make sense to sensor and actuator manufacturers, hardware engineers that don't need to switch to delivering software at scale.
It's clear that AISS is a key component of Microsoft's instantiation of Nadella's ambient intelligence. By using technologies like Hadoop and the rules engine built into the Azure service bus, Microsoft is building a fast analytics platform that can work with both streamed and historic data, that's also capable of delivering an automated response to that data: a response that can be displayed on user dashboards, or that can be used to trigger an action on a controller at the other end of an IP address.
Big data needs lots of data to generate insights, and that's what the Internet of Things offers. If those insights are to become intelligence, then there's still a lot of work to be done -- and at scale. It's work that needs a cloud to work, and Microsoft is betting that Azure is the right cloud for the job.