Google Inc. today issued a big challenge to its rivals in cloud computing by opening up access to what has been described as the world’s largest database.

The company is launching Cloud Spanner Beta, providing software developers with a database service available through Google Cloud that the search giant already uses to run its massive AdWords advertising system and Google Play app and media store.

Google’s Spanner is a relational database, the kind that stores data in familiar related rows and columns. What sets it apart from many rivals, including IBM Corp.’s DB2, the Oracle Corp. Database, Microsoft Corp.’s SQL Server and the open-source MySQL, is that it can scale up globally across hundreds of data centers and millions of machines yet act as a single database to keep data consistent in near-real-time.

Essentially, Spanner can handle a ridiculous number of transactions around the world at once, keeping them in order without having to replicate the data in a lot of data centers. That can be costly and can cause delays in recording the data and providing access to high-velocity applications such as stock transactions and automated teller machines.

Google says Cloud Spanner offers both data consistency across multiple databases while also making the data highly available. In short, Cloud Spanner does for distributed transaction processing, said Ovum Ltd. Principal Analyst Tony Baer, what Hadoop software did for big data analytics.

Databases are a crucial foundation for cloud computing, because all applications are built atop them. One reason Oracle remains dominant in databases is that it’s very hard to move data off one brand of database to another. If Google can woo enterprises to Cloud Spanner, it may be able to keep them as customers for a long time — customers that will end up using many other Google services as well.

“This is the battleground for the next phase of cloud adoption,” said Peter Burris, chief research officer at Wikibon, the analyst group owned by the same company as SiliconANGLE. And not just for the cloud giants. Last week, for instance, the cloud infrastructure firm Stratoscale said it’s buying database-as-a-service startup Tesora Inc., as well as launching its own relational database service compatible with Amazon Web Services.

A credible database offered as a service is crucial to Google’s efforts to expand its cloud computing business, which in turn is key to keeping Google in the hearts and minds of software developers — not to mention potentially providing another multibillion-dollar business beyond advertising. Google lately has gained some ground in the cloud, winning some big deals such as a five-year, $2 billion deal for cloud services to soon-to-go-public Snap Inc.

Long game

But according to Synergy Research Group Inc., Google Cloud still trails far behind Amazon Web Services and Microsoft Azure. That’s one reason Google has been working for much of the past year or more, according to one report in The Information, to commercialize Spanner. Analysts briefed on the launch said they think it could be a potent competitive differentiator for Google versus AWS, Azure and even database king Oracle Corp. “Nobody else has Spanner,” said Gartner Inc. analyst Nick Heudecker. “And Google has the resources to play the long game.”

Regardless of the database’s power, in fact, this will be a very long game for Google. For one, it’s not yet clear how well Spanner will work for use cases outside Google’s own. Making databases work at huge scale requires lots of tweaks and black magic particular to each company’s applications.

Some early users of Cloud Spanner say it’s working for them. Among the customers Google trotted out is Quizlet Inc., which provides an online learning tool to some 20 million teachers and students. As it has grown, its MySQL database has had to be split up, or sharded, reaching a query capacity limit. Cloud Spanner, said Quizlet Platform Lead Peter Bakkum, “could dramatically simplify how we manage our databases.”

The bigger challenge will be very large enterprises that already have a workable system, one that is reliable and meets compliance and government regulations. They simply won’t move that system wholesale over to a new database no matter how good it looks. It’s far too risky, said Burris, since failures in these systems can mean people go to jail. That’s why “moving to a new transaction database is a long slog,” said Baer, one that can take five years or more.

On the other hand, some of the most demanding large enterprises, such as financial institutions, are also among the most forward-looking when it comes to new technologies, especially those that can give them a time or cost advantage. Coupled with some customers already trying out Cloud Spanner, said Heudecker, “this certainly makes their cloud value proposition clearer.”

Photo: © Can Stock Photo/nevarpp