The latest addition to Google Cloud Platform, Cloud Bigtable is a managed NoSQL database solution accessible through the HBase API, making it easy for developers already using HBase and Hadoop to integrate Cloud Bigtable into their applications and move their data into and out of Google's platform.
Cloud Bigtable also plays nicely with other Google offerings, including BigQuery and Cloud Dataflow, which allow developers to process and analyze large sets of data. Google says that Cloud Bigtable clusters can be created or reconfigured in seconds, and storage is scalable, eliminating the need for users to estimate their storage requirements. Data transmitted to and from Cloud Bigtable is encrypted while in transit and at rest. According to Google's benchmarking, Cloud Bigtable offers more than double the write throughput per dollar and exponentially lower read/write latency as compared to HBase and Cassandra.
By bringing the technology Google has used internally to run its most successful services for more than a decade, the search giant is targeting a variety of big data users, including companies in the financial services, energy, biomedical and advertising technology industries.
In an effort to drive adoption of Cloud Bigtable, Google has established a number of partnerships with services firms and solutions providers that have integrated Cloud Bigtable into their platforms. It has also created documentation and code detailing how developers can set up an HBase server with REST API that can communicate with Cloud Bigtable.
Google offers Cloud Bigtable on the same pay-as-you-go basis as most Cloud Platform offerings. Each Cloud Bigtable node, which can deliver 10,000 queries per second and 10 MB per second of data transfer, costs 65 cents per hour. At least three nodes are required per Cloud Bigtable cluster. There are also fees for storage and some data transfer.
Cloud Bigtable, which is in beta, is similar to Google Cloud Datastore, another NoSQL database service, but is designed for more demanding use cases. It competes with offerings like Amazon's DynamoDB, which also promises developers a highly scalable NoSQL database appropriate for low-latency applications.