Google announced the expansion of Cloud BigTable's replication capabilities in Beta, providing customers with the flexibility to make their data available across a region or worldwide. Furthermore, the enhancements will allow customers to create a replicated cluster in any zone at any time.
Google's Cloud BigTable is a globally distributed wide-column NoSQL service on the Google Cloud Platform (GCP). The service can scale from gigabytes to petabytes, and provide high-performance throughput and low latency response times similar to other Google services such as Search, Google Maps, and YouTube.
In August of last year, Google made a regional replication for the service generally available and is now expanding the replication capabilities globally in beta. Moreover, the new capabilities will simplify the distribution of data, with closer access to data, and provide a single location for aggregated data from sources worldwide such as IoT Sensor data. The ability to enhance availability and durability using multiple regions is also provided. As Sandy Ghai, product manager at Google, explains in the announcement in a blog post:
Every cluster in a replicated instance accepts both reads and writes, providing multi-primary replication with eventual consistency. You can set up replication by adding one or more Cloud Bigtable clusters, whether on the same continent or halfway around the world.
Source: https://cloud.google.com/blog/products/databases/go-global-with-cloud-bigtable
With the new replication capabilities, users can configure a replication topology using any zones where Cloud Bigtable is available, or add clusters in additional regions to an existing instance without any downtime. Furthermore, the replication model will allow users to reconfigure their instance's replication topology at any time by allowing the addition or removal of clusters for any existing instance, even if they are writing data to that instance. Ghai explains in the same blog post on what happens when users add a cluster to an existing instance:
- First, all existing data will be bulk-replicated from the existing cluster to the new one
- Then, all future writes to any cluster will be replicated to all other clusters in the instance
Note that all tables within an instance are replicated to all clusters, and users can monitor replication progress for each table via the Tables list in the GCP Console.
Comparable NoSQL services for Cloud Bigtable are Cosmos DB from Microsoft and DynamoDB from Amazon - all having replication capabilities. Moreover, Cosmos DB and DynamoDB offer similar capabilities as Cloud Bigtable, however there is a significant difference in supporting database models and APIs for accessing data.
Lastly, Google increased the availability of Cloud BigTable to 17 GCP regions including Brazil, India, Australia, and China with more coming soon. Pricing details are available on the pricing page.
