VIDEO
HIGHLIGHT
Log InSign up
Cloud Bigtable
SummaryTranscriptChat

Cloud Bigtable

New Section

If you don’t require transactional consistency, Cloud Bigtable is a fully managed NoSQL database that offers petabyte-scale storage and low latency. It seamlessly scales for throughput and adjusts to specific access patterns. Cloud Bigtable powers many of Google’s core services, making it suitable for operational and analytical applications.

Cloud Bigtable Features

  • Cloud Bigtable is a fully managed NoSQL database with petabyte-scale storage and low latency.
  • It seamlessly scales for throughput and adjusts to specific access patterns.
  • Powers many of Google’s core services like Search, Analytics, Maps, and Gmail.
  • Suitable for both operational and analytical applications such as IoT, user analytics, financial data analysis, and machine learning.
  • Integrates easily with popular big data tools like Hadoop, Cloud Dataflow, and Cloud Dataproc.
  • Supports the open-source industry standard HBase API for easy development.

Data Structure in Cloud Bigtable

  • Cloud Bigtable tables store data in massively scalable tables composed of rows and columns.
  • Each row is indexed by a single row key.
  • Columns related to one another are grouped together into column families.
  • Each column is identified by a combination of the column family and a unique column qualifier name within the family.
  • Each row/column intersection can contain multiple cells or versions at different timestamps.

Example Use Case: Social Network

Let's highlight some key points about using Cloud Bigtable for a hypothetical social network:

  • The table contains one column family called "follows" with multiple column qualifiers used as data. This design choice takes advantage of the sparseness of Cloud Bigtable tables.
  • The username serves as the row key to evenly spread data access across the entire table.

Architecture Overview

Cloud Bigtable Architecture

  • Processing is handled separately from storage in Cloud Bigtable.
  • Cloud Bigtable tables are sharded into tablets (similar to HBase regions) stored on Google's file system, Colossus, in SSTable format.
  • SSTables provide a persistent, ordered immutable map from keys to values.
  • Cloud Bigtable adjusts indexes based on access patterns to distribute workload evenly across nodes.

Choosing Cloud Bigtable

Consider using Cloud Bigtable if you:

  • Need to store more than 1 TB of structured data.
  • Have a high volume of writes and require read/write latency of less than 10 milliseconds with strong consistency.
  • Require compatibility with the HBase API.

New Section

Firestore is recommended for storage services that scale down well. It offers flexibility and ease of use for smaller-scale applications.

Scaling Down with Firestore

  • If you don't need the specific requirements mentioned earlier, consider using Firestore.
  • Firestore is suitable for storage services that scale down well.
  • Offers flexibility and ease of use for smaller-scale applications.

The smallest Cloud Bigtable cluster consists of three nodes.