The following table lists the basic concepts of AnalyticDB for PostgreSQL.

Term Description
Massively Parallel Processing (MPP) A distributed shared-nothing computing architecture. MPP uses a large number of nodes that do not share resources with each other to perform parallel computing and improve performance. A node is known as a data partition in AnalyticDB for PostgreSQL.
compute node The unit for allocating cluster resources in AnalyticDB for PostgreSQL. An instance is composed of multiple compute nodes. When the number of compute nodes increases, the storage capacity scales out but the query response time does not change. A compute node is a unit of computing resources that includes fixed CPU cores, memory, and storage. Each compute node type contains a data partition in the MPP architecture.
number of compute nodes The number of compute nodes purchased for a cluster instance. A single instance supports up to 4,096 nodes. The storage space and compute resources of cluster instances increase linearly as the number of compute nodes increases.
data partition A unit for computing and storing data in the MPP architecture. Table data is stored in different data partitions by partition key.