IT teams deploy cloud database services to hand off some of the management and operations tasks to a service provider so they can focus on actually putting that data to use. Cloud databases can also provide scalability, failover support and potential cost savings.
Google, like its competitors, offers different cloud database options for different needs. Explore seven popular Google Cloud database services and see if any fit your workload requirements.
Google Cloud SQL. This managed relational database is compatible with PostgreSQL and MySQL. Google also added support for SQL Server to attract users with Windows workloads. To ensure users’ databases are flexible, highly available and consistent, Cloud SQL automates replication, backups, encryption and failover.
Users can connect to this service from Compute Engine, Google Kubernetes Engine, App Engine and BigQuery. Pricing for this Google Cloud database option, which includes a per-second billing model, varies based on instance type.
Google Cloud Spanner. Though it’s a distributed, relational database service, Cloud Spanner eliminates some of the consistency and availability tradeoffs typically associated with this database type. That’s because while it provides relational database semantics like schemas, SQL queries and ACID transactions, it also incorporates horizontal scaling capabilities often associated with NoSQL databases. This Google Cloud database option is designed to support global online transactional processing and can scale to thousands of nodes distributed across global regions.
Pricing is based on nodes, networking and storage. Nodes are priced on an hourly basis, while networking and storage are charged on a monthly basis. Storage billing is based on the average amount of data in a Cloud Spanner table and other secondary indexes, while networking costs depend on the amount of bandwidth used during that month.
Google Cloud Bigtable. This managed columnar database can scale to billions of rows and columns, which enables users to store petabytes of data. Cloud Bigtable is best suited for storing large amount of single-keyed data with low latency. Users can store different types of data, including time-series, marketing, financial, IoT and graph data.
Cloud Bigtable also integrates with popular big data tools such as Hadoop and supports the open source HBase API. Cloud Bigtable users are charged for storage and network bandwidth, as well as the instance type — production or development — and the total number of nodes in a cluster.
Google Firebase Realtime Database. This hosted NoSQL database is primarily used for web and mobile application development. Google Firebase Realtime Database provides data in real time for users, in JSON format, which is then stored in a large JSON tree. Thus, this product is best suited for handling simple data, as opposed to data in large amounts or in a hierarchical manner.
This Google Cloud database option can be maintained offline through a local cache, even if an application is not connected to the internet. Google Firebase Realtime Database automatically updates and syncs once the app reconnects, which makes it possible for the application to function if the network is slow or disconnected.
While this offering is reliable and fast, it’s limited to the availability zones of one region. In terms of pricing, there is a free Spark Plan for limited storage and capabilities or the Blaze Plan, which has pay-as-you-go pricing based on the amount of gigabytes used per month.
Google Cloud Firestore. Although this database is a part of the larger Firebase portfolio, it is not the same as Firebase Realtime Database. Firestore is a more flexible and scalable NoSQL database option, and it’s suitable for more complex application development. Firestore also offers more features and functionality for newer or demanding workloads.
Firestore has an offline mode, similar to Firebase Realtime Database. However, it can also integrate with other Google Cloud services and open source technologies, which makes it more versatile and flexible when designing an application. Firestore also offers the reliability of native multi-region services, rather than being limited to a single region. Billing for this Google cloud database option depends on the sum of reads, writes and deletes performed, as well as the total amount of storage and network bandwidth used.
Google Cloud Memorystore. Unlike Google’s other databases, Memorystore is a managed Redis service. Redis is an open source key-value database that is primarily used for cache management and web application speed. By using Memorystore for Redis, users can build application caches that meet Redis protocol, which enables a streamlined migration process.
This frees up developers by automating normally tedious tasks, such as patching and failover. Its key features include security, monitoring, migration, high availability and more. Users are charged per gigabyte per hour, based on three components: service tier, provisioned capacity and region.
Google BigQuery. Although this Google service is not explicitly a database service, it is useful for businesses looking for a serverless, scalable, cost efficient data warehouse. BigQuery enables users to closely analyze data using a SQL-like syntax for their organization, without having to manage the service itself. It provides real-time data and insights to predict business outcomes.
There are also BigQuery variants, including: BigQuery ML to build and operationalize ML models, BigQuery BI Engine for the analysis of complex datasets and BigQuery GIS that combines BigQuery’s serverless architecture with geospatial analysis. This data warehouse also offers storage and compute separation, automatic backup and easy restore, as well as big data ecosystem integration and several other features. Pricing is broken down by storage — charges vary depending on modification activity — and query costs. Users can choose between on-demand or flat-rate pricing models.