Google Compute Services

Nearline is a way to use Google Cloud Storage to backup and archive data – the kind you wouldn’t necessarily think of in a database, which can only be accessed once, by one user, and usually no more than once a month. Google calls this model “cold storage,” and it adapts its pricing model to account for this low level of use, with the goal of making Nearline a more attractive option for purposes like system backups.
Anthos, announced last April, is a GCP system for organizing and maintaining applications that may be centered around Google, but which may use resources from AWS or Azure (“Multiple Cloud Services”). Think of an app that Google hosts its own code base, but borrows AI functionality from AWS and stores its records in an Azure object vault.
BigQuery is a Google Cloud Storage that uses a data storage system designed for very large amounts of highly distributed data, enabling SQL queries to be executed across multiple databases with different architecture levels. Instead of the traditional row-based and record-oriented SQL relational database indexing, BigQuery uses a vertical storage system in which the components of records are stacked on top of one another and flowed into a parallel storage system. Such an organization proves useful in analytics applications, which gather broad statistics about the often simple and general relationships between data elements.
Cloud Bigtable (formerly BigTable) is a highly distributed data system that organizes related data into a multi-dimensional aggregation of key/value pairs, based on the extensive storage system created by Google for its own use in storing search indexes. Such aggregation is easier for analytics applications to manage than a very large index to a huge relational database with multiple tables whose records must be joined at the time of the query.
Cloud SQL (not yet ready for public consumption) hosts much more traditional relational database tables and indexes, using a GCE instance that develops itself to meet database performance requirements.
Cloud translation, text-to-speech, and speech-to-text, as their names suggest, from Google’s current ability to manage spoken and written language, for use in dedicated applications.
Apigee is a modeling system for producing and managing APIs – service calls for server-based functions, using the web as a communication medium. Apigee user can design, test, and deploy mechanisms for their existing web applications to be discoverable using APIs, and monitor how web users use these API calls for their own purposes.
Istio is an interesting type of “phone book” for modern, scalable applications that are distributed as individual components called microservices. The adjacent classic app knows where all of its functions are located; A microservice-based application needs to be aware, via a service network. Istio was originally developed as a service network through an open-source partnership consisting of Google, IBM, and Lyft.
Cloud Pub / Sub (publish and subscribe) is a mechanism that replaces the message queues that middleware used during the previous era of client/server applications. For applications designed to collaborate without explicitly communicating with each other (“asynchronously”), Pub / Sub acts as a kind of post office for events.
Cloud AutoML is a suite of services targeted at enabling applications to take advantage of machine learning – to discover perceived patterns across large amounts of data, and to use those patterns within the software.

This is not a complete list of Google Compute services, although it does give you the key entries. In fact, some of the company’s many services are applications or reconfiguration of other services – ways to use a service that performs a broad function, for a more specific purpose.