Bekker's Blog

Blog archive

Nadella Touts Rapid Rise of Microsoft's Cosmos DB

Just a year after launching Cosmos DB as an Azure service at the 2017 Build developer conference, Microsoft CEO Satya Nadella says it's the fastest-scaling database he's ever seen.

"In less than a year, Azure Cosmos DB, the first globally distributed [and] multi-model database, exceeded $100 million in annualized revenue," said the Microsoft CEO during the company's Q3 earnings call last week.

While annualized revenue doesn't mean Microsoft has pulled down $100 million on Cosmos DB yet, it does mean that it's recently ramped up to a pace of more than $8.3 million a month.

"I've been around databases for a long time. I've never seen a product that's gotten to this kind of scale this quickly," Nadella said, according to the Seeking Alpha transcript of the call.

Replacing Microsoft's previous NoSQL offering DocumentDB, and competitive against Amazon Web Services (AWS) DynamoDB and Google Spanner, Cosmos DB boasts several key selling points.

As a base service of every Azure datacenter and region, Cosmos DB is globally available and allows developers to click on a map to add or delete geographic regions even while their Cosmos DB-based applications are running. That worldwide distribution allows for both massive scale and minimal latency for users anywhere.

Cosmos DB includes support for multiple data models, including document, graph, key-value, table and column-family, and has APIs for SQL, MongoDB, Cassandra, Gremlin and Table.

Microsoft also stands behind Cosmos DB with guarantees, including a four-nines uptime service-level agreement, end-to-end latency times in the low millisecond range and zero-data loss in the case of regional failures.

During the investor call, Nadella positioned Cosmos DB as a potential cornerstone of future data projects on Azure, especially as artificial intelligence efforts drive demand for more and more data. "This AI era is mostly first a data era. And that's where I think the opportunity lies," Nadella said. "Cosmos DB happens to be one of the best database products to be able to capture the signals that you want around your customers from a variety of different sources."

How many customers would be required to drive more than $100 million in annualized revenue is not clear. From the earliest discussions at Build last year, it was evident that even a few customers can drive massive data volumes. At that show, Microsoft Executive Vice President of Cloud and Enterprise Scott Guthrie said early adopters, including Jet.com, were already using Cosmos DB to the tune of 100 trillion transactions per day.

Andrew Brust, founder and CEO of Blue Badge Insights and a co-chair of the 1105 Media Visual Studio Live! conference series, is seeing broad interest in Cosmos DB in the database community.

"My own observation of Microsoft tech influencers who have years -- or decades -- of experience with conventional database technology like SQL Server (and therefore Azure SQL Database), is that they are kicking the tires on Cosmos DB and still working to understand how its pricing will work for them in production," Brust said in an e-mail interview.

"The use case for massive, global-scale Web properties employing a database like Cosmos DB is pretty clear. For the long tail of enterprise applications with smaller, less concurrent and less geographically dispersed user bases, the adoption of Cosmos DB is more disruptive and less straightforward," Brust said. "This will likely get better with time, and the pricing models for Cosmos DB will likely become more compelling and more easily understood, too. Amazon's DynamoDB has been out for many years and is explicitly integrated with a number of other AWS services so, clearly, Cosmos DB isn't on a level playing field with it yet. This, too, will change over time."

Brust says a key point to remember about Cosmos DB is that it's based on the same technology Microsoft has used for its own cloud services for years.

"The company's skin is in the proverbial game, with very high stakes. If it's worked for them, it's going to work really well for their customers. What's left is more fit and finish on pricing, marketing and rationalizing the service with other components in the Azure data and analytics stack," Brust said.

Posted by Scott Bekker on April 30, 2018 at 10:39 AM