Google launches Cloud Spanner scalable database
Wed 15 Feb 2017
Google is releasing a version of its internal, NoSQL-style scalable database Cloud Spanner to public beta. The format is intended to straddle the needs of those who have until now had to choose between the transactional consistency of traditional linear databases and the more elastic scaling and data distribution of NoSQL database formats.
Offering a proprietary format closes the circle for Google’s cloud offerings, as the company will now be able to offer the core db format in concert with its connectivity, compute and storage outings, potentially freeing it from specious patent lawsuits, and from licensing fees for more established market formats.
Cloud Spanner is said to feature data-layer encryption, disaster protection and high availability, and its nodes will be chargeable on a per-hour basis, with one hour as the minimum, coming to $0.90 USD per node per hour. Storage for tables and secondary indexes is sold in gigabytes averaged over a month at $0.30 USD per GB monthly.
Incoming traffic is free on a per gigabyte basis, with exiting traffic chargeable: up to 1TB launches at $0.19 per GB in Australia, $0.23 in China (excl. Hong Kong), and $0.12 elsewhere. Between 1-10TB the same respective regions will be charged $0.18, $0.22 and $0.11 per GB. Over 10TB the tariff comes to $0.15, $0.20 and $0.08.
Cloud Spanner Product Manager Deepti Srivastava says that the format will support ‘Atomicity, Consistency, Isolation, Durability” (ACID) in transactions and preserving SQL semantics, “without giving up horizontal scaling and high availability.’ and adds that 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 announcement notes that software supply chain and retailer JDA was an early adopter of the Cloud Spanner-based platform in 2015, and emphasises that Cloud Spanner is a complementary product with respect to its Cloud Datastore and Cloud Bigtable.
It’s claimed that the format offers horizontal scaling without the need for relational>NoSQL migrations, and that it can scale out RDBMS solutions without complex sharding or cluster issues.