Everyone sees how useful data is to their business these days, observes T.K. Rangarajan, who runs the teams creating Microsoft's core database platforms. In fact, they're beyond enthusiastic.
"The world is drunk on data," he says. "Everyone has seen enough evidence that with data you can get great insights, with data you can see the effect of recommendations, like people liking your stuff on Facebook. Now business leaders are looking at these sorts of examples and interpreting them for their own business, and they're asking 'why can't I have that?' There is a tremendous deep desire on the part of businesses to achieve their objectives with greater, deeper, more intensive cultivation of data."
Getting the data isn't the problem. "There's a proliferation of data," he points out. "Every human interaction with a digital system – every time you come into contact with a company, every time you send an email or a text or a tweet – they're all opportunities for somebody to learn about what you want, what you think, what you value, what's popular. Then there's all these sensors measuring temperature and humidity… Data is produced in all sorts of places."
Coping with the data flood
But dealing with that isn't easy, because so far there hasn't been an easy way to take what developers and IT administrators know about building business systems and use that for running big data systems. The new Microsoft data services change that, because they use familiar SQL options – but integrate with Hadoop systems, as well as Microsoft's own cloud analytics and machine learning services.
"Our approach for developers is to bet on standards so we can leverage the familiar tools, technologies and frameworks they're used to," says Rangarajan. "The world as we know it for running businesses is relational databases – we know how to run businesses on it. SQL is a well-trodden path. We know how to do business intelligence, how to load data into a warehouse, how to optimise your sales force every quarter; we know how to do that."
"From a data point of view we make it easy for developers to extract value from any type of data, whether it's richly structured, JSON, flat relational transactional or flat relational data warehouse, or completely unstructured Hadoop. You can analyse that data, you can do machine learning on any data that you move into Azure from anywhere.
"We want to make it easy for developers to get meaning out of this data, to put it together and use visualisation tools for human insights or machine learning tools for automated decisions. We want to support the ways insight is shared and creates value."