A spanner in the works: Google's cloud database hits beta, gets prices

Ye canna' hand a man a grander (Cloud) Spanner

Google's close to plugging a long-standing gap in its public cloud, with its Cloud Spanner distributed relational database hitting public beta.

In January, we noted that Cloud Spanner, first detailed in a 2012 white paper, had landed as alpha in 2014 but was yet to become a commercial offering.

The beta announced February 14 brings that a step closer.

Cloud Spanner's product manager Deepti Srivastava writes that it's a “have cake and eat it” proposition, supporting “Atomicity, Consistency, Isolation, Durability” (ACID) in transactions and preserving SQL semantics, “without giving up horizontal scaling and high availability”.

The Chocolate Factory will handle high availability and disaster protection for users, and promises data-layer encryption, identity and access management, and audit logging.

Pricing – which Google told us in January is the detail in which lives the devil – is fairly straightforward:

  • Cloud Spanner nodes are charged on a per-hour (minimum one hour) basis, according to the maximum number of nodes that exists in each hour (US$0.90 per node per hour);
  • Storage (in GB, for tables and secondary indexes) is averaged over a month, and charged by a monthly rate ($0.30 per GB per month); and
  • Bandwidth in GB is charged by traffic exiting Cloud Spanner, while ingress traffic is free.

Egress traffic within a region is free, and between regions in the USA is a cent per GB. Internationally, traffic charges are more complex:

GB per Month (egress only) Australia China (excluding Hong Kong) Rest of World (including Hong Kong)
Up to 1 TB $0.19 per GB $0.23 per GB $0.12 per GB
1 TB to 10 TB $0.18 per GB $0.22 per GB $0.11 per GB
10 TB + $0.15 per GB $0.20 per GB $0.08 per GB

(Cue outrage from Australian readers suffering from the country's high Internet transit costs.)

Srivastava adds Cloud Spanner “supports distributed transactions, schemas and DDL statements, SQL queries and JDBC drivers and offers client libraries for the most popular languages, including Java, Go, Python and Node.js.”

Google's posted a product video to YouTube, below. ®

Youtube Video


Biting the hand that feeds IT © 1998–2017