Related topics

Microsoft parts Azure cloud, reveals NoSQL doc database

We're not in a relational world anymore, Dorothy

Cloud security image

Microsoft has slipped out DocumentDB for Windows Azure, the company’s first-ever non-relational database – and its first new database product since SQL Server.

DocumentDB is a complete departure from Microsoft’s relational roots, being a schema-free, NoSQL offering built entirely for consumption as a service on its cloud.

Microsoft said it’s going NoSQL to enable “new scenarios” on Windows Azure – meaning mobile and web.

The company’s recently installed server and tools chief Scott Guthrie said Microsoft has had DocumentDB in development for a year now.

He claimed instances of DocumentDB now running on Windows Azure that are hundreds of terabytes in size and processing millions of complex queries per day.

DocumentDB throws a curve ball to the document-oriented, non-relational stores Microsoft had lured into running on Windows Azure to make it more interesting and relevant.

MongoDB has been available on Windows Azure since December 2011 thanks to some intense collaboration between the engineering teams of both companies to iron out the wrinkles. New MongoDB instances for Azure were rolled out as recently as April. It’s also possible to get CouchDB up and running on Microsoft’s cloud.

Microsoft is bowling its own document-oriented NoSQL at both the diehard Windows shops going web and mobile and the previously wouldn’t-touch-Microsoft-with-a-10-foot-pole types deep into open technologies.

Companies like MongoDB and Couchbase, backing CouchDB, are tiny compared to Microsoft but have proven customers – the kinds of big names Microsoft likes.

Mongo claims a long list in various sectors with a strong showing in media, while Couchbase claims eBay, Orbitz and Salesforce as customers.

Like so many recent innovations, Microsoft is coming from behind on these. Firing up the open-standards flare, Guthrie said: “DocumentDB has made a significant bet on ubiquitous formats like JSON, HTTP and REST – which makes it easy to start taking advantage of from any web of mobile applications.”

He also hit them on scalability, saying those DocumentDB instances are doing so with “predictable performance of low, single digit” millisecond latency. Microsoft is divvying up DocumentDB into capacity units, added or removed through the Azure portal or REST API.

Unveiling DocumentDB, Guthrie also bowled out development kits for .NET, Node.js, JavaScript and Python and a new Windows Azure Search service and API Management REST API.

Microsoft hopes devs will use DocumentDB with the REST management API to run things like products and users, managing things like subscriptions and billing.

DocumentDB is a NoSQL store, meaning it’s schema free and lets you store JSON documents while querying them using a document-oriented SQL query language.

Microsoft is best known for its SQL Server relational database, a multi-million dollar business responsible for rapid growth in its server and tools unit. Microsoft developed SQL Server with Sybase and Ashton-Tate delivering 1.0 in 1989 orignally for use with dBase but this didn't work out.

SQL Server was put on Windows and on IBM's OS/2 but Redmond then really still saw itself as a PC operating system and apps company. It only got serious on the database once its relatioship with both Aston-Tate and Sybase had finished and it was free to do with the code what it wanted. Looking to grow beyond the PC, Microsoft released SQL Server 4.21 for its fledgling server operating system, Windows NT, in 1993.

It was a huge success. Recently, SQL Server is credited by Microsoft as the chief reason all server revenues grew 11 per cent during the company’s last fiscal year, to $1.7bn. SQL Server was released as the company’s answer to DB2 and Oracle on Windows.

NoSQL databases have been pushed first as the alternative and now as a complement to relational databases like SQL Server. They store data in structured rows and columns for managing and querying huge volumes of unstructured data. They do this using open tools and language, which means you don't necessarily need to speak pure relational. ®

Sponsored: Designing and building an open ITOA architecture