The Schwartz Report

Blog archive

Amazon Seeks To Offer SQL Alternative with Aurora Cloud Database

When Amazon Web Services announced Aurora as the latest database offering last week, the company put the IT industry on notice that it once again believes it can disrupt a key component of application infrastructures.

Amazon debuted Aurora at its annual AWS re:Invent customer and partner conference in Las Vegas. Amazon said the traditional SQL database for transaction-oriented applications, built to run on monolithic software and hardware, has reached its outer limits. Amazon Web Services' Andy Jassy said in the opening keynote address that the company has spent several years developing Aurora in secrecy.

Built on the premise that AWS' self-managed flagship services EC2, S3 and its Virtual Private Cloud (VPC) are designed for scale-out, service-oriented and multi-tenant architectures, Aurora removes half of the database out of the application tier, said Anurag Gupta, general manager of Amazon Aurora, during the keynote.

"There's a brand new log structured storage system that's scale out, multi-tenant and optimized for database workloads, and it's integrated with a bunch of AWS services like S3," said Gupta, explaining Aurora is MySQL-compatible. Moreover, he added, those with MySQL-based apps can migrate them to Aurora with just several mouse clicks and ultimately see a fivefold performance gain.

"With Aurora you can run 6 million inserts per minute, or 30 million selects," Gutpa said. "That's a lot faster than stock MySQL running on the largest instances from AWS, whether you're doing network IOs, local IOs or no IOs at all. But Aurora is also super durable. We replicate your data six ways across three availability zones and your data is automatically, incrementally, continuously backed up to S3, which as you know is designed for eleven nines durability."

Clearly Amazon is trying to grab workloads that organizations have built for MySQL but the company is also apparently targeting those that run on other SQL engines that it now hosts via its Relational Database Service (RDS) portfolio including Oracle, MySQL and Microsoft's SQL Server.

Aurora automatically repairs failures in the background recovering from crashes within seconds, Gupta added. It can replicate six copies of data across three Availability Zones and backup data continuously to S3. Customers can scale an Aurora database instance up to 32 virtual CPUs and 244GB of memory. Aurora replicas can span up to three availability zones with storage capacities starting at 10GB and as high as 64TB.

Gupta said the company is looking to price this for wide adoption, with pricing starting at 29 cents for a two-virtual CPU, 15.25-GB instance.

The preview is now available. Do you think Amazon Aurora will offer a viable alternative to SQL databases?

Posted by Jeffrey Schwartz on 11/17/2014 at 12:32 PM


comments powered by Disqus

Subscribe on YouTube