To understand core concepts when you plan, deploy, and manage
AlloyDB Omni, review the following definitions of architecture,
Kubernetes orchestration, caching, and performance terms.

## A

AlloyDB AI
:   A suite of features included with AlloyDB Omni that let you build enterprise generative AI applications.

AlloyDB Omni using the container orchestrator
:   An extension to the Kubernetes API that lets you manage AlloyDB Omni in most CNCF-compliant Kubernetes environments.

## D

database cluster
:   A group of interconnected database instances that work together to provide high availability, scalability, and fault tolerance for database applications.

disk cache
:   Accelerates data retrieval in AlloyDB Omni installations in which the active working set of data doesn't fit in memory.

## I

instance
:   A representation of your data housed in one or more virtual machines during the runtime of a computer program, and is the main unit of cloud computing resources in AlloyDB Omni and SQL.

## L

latency
:   The amount of time it takes to process a single request, usually measured in milliseconds, seconds, or sometimes larger time amounts for complex reporting and analytics use cases. Also known as response time.

## S

scheduling
:   In Kubernetes, a process for matching new database pods to nodes to balance node distribution across the Kubernetes cluster and help optimize performance.

## T

throughput
:   The amount of work the database can process in a given amount of time. Throughput is usually expressed as transactions per second (tps) or transactions per minute (tpm).

## V

vCPU
:   Virtual central processing unit. A single hardware thread as observed by the operating system and available to be used for execution by an AlloyDB Omni database instance.