Google Cloud makes it cheaper to run smaller workloads on Bigtable

Cloud Bigtable has long been Google Cloud’s fully managed NoSQL database for massive, petabyte-sized analytical and operational workloads. At $0.65 per hour and node, it was never a cheap service to run, especially because Google Cloud enforced a minimum of three nodes per cluster for production workloads. Today, however, it is changing that, and you can now run Bigtable production workloads on just a single node.

“We want Bigtable to be an excellent home for all of your key-value and wide-column use-cases, both large and small,” Google Cloud Bigtable product manager Sandy Ghai said in today’s announcement. “That’s true whether you’re a developer just getting started, or an established enterprise looking for a landing place for your self-managed HBase or Cassandra clusters.”

With this, Google Cloud is now also enabling the ability to use replication for higher availability for these small clusters, as well as the ability to easily switch a one-node development instance to a one-node production instance as needed. In addition, the service’s SLA now also covers all Bigtable instances, no matter their size.

It’s interesting to see Google Cloud make this push for bringing smaller workloads onto Bigtable, especially given the organization’s current focus on large enterprise customers and their specific needs. But the company that only needs a single node today could easily be the one that needs massive clusters in the future and Bigtable’s minimums have always represented somewhat of a barrier to entry for smaller companies — and once a company places its bets on a given database service, it’s not likely to switch anytime time.