Chapter 7. Cloud Bigtable: large-scale structured data
This chapter covers
- What is Bigtable? What went into its design?
- How to create Bigtable instances and clusters
- How to interact with your Bigtable data
- When is Bigtable a good fit?
- What’s the difference between Bigtable and HBase?
Over the years, the amount of data stored has been growing considerably. One reason is that businesses have become more interested in the history of data changes over time than in a snapshot at a single point. Storing every change to a given value takes up much more space that a single instance of a value. In addition, the cost of storing a single byte has dropped significantly. Following this practice has led to engineering projects focused on discovering more uses for all of this just-in-case data such as machine learning, pattern recognition, and prediction engines.
These new uses require storage systems that can provide fast access to extremely large datasets, while also maintaining the ability to update these datasets continuously. One of these systems is Google’s Bigtable, first announced in 2006, which has been reimplemented as the open source project Apache HBase. Based on the success of HBase, Google launched Cloud Bigtable as a managed cloud service to address the growing need for these large-scale storage systems. Let’s explore what Bigtable is and dig into some of the technical details that went into building it.