> For the complete documentation index, see [llms.txt](https://mariadb.com/docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://mariadb.com/docs/gridgain/gridgain-8/reference/glossary.md).

# Glossary

Definitions of the core GridGain 8 terms used throughout this documentation — clusters, nodes, caches, partitioning, persistence, and more.

Core terms used throughout the GridGain 8 documentation.

**Affinity colocation** Placing related data (and, optionally, computations) on the same node so that operations such as SQL joins run without moving data across the network. See [Affinity Colocation](/docs/gridgain/gridgain-8/architecture/data-distribution/affinity-colocation.md).

**Atomicity mode** The consistency guarantee configured on a cache — `ATOMIC` for the highest performance, or `TRANSACTIONAL` for full ACID transactions.

**Baseline topology** The set of server nodes that persist data and are expected to be present in the cluster. See [Baseline Topology](/docs/gridgain/gridgain-8/architecture/cluster/baseline-topology.md).

**Binary object** GridGain's platform-independent representation of cached objects, allowing data written by one language client to be read by another without shared classes.

**Cache** The primary data structure that holds key-value entries; equivalent to a SQL table. See [Cache vs. Table](https://mariadb.com/docs/gridgain/gridgain-8/reference/pages/OUSGdNs5HGVTk0HR0E50#cache-vs.-table).

**Client node** A node that connects to the cluster to run application logic but does not store data or contribute to computations by default.

**Cluster** A group of interconnected nodes that pool their memory and CPU to store and process data.

**Data region** A configurable block of memory that caches are assigned to, with its own size limits and persistence settings. See [Configuring Data Regions](/docs/gridgain/gridgain-8/gridgain8-development/caches-and-memory/memory-configuration/data-regions.md).

**Discovery** The mechanism by which nodes find and join each other to form a cluster. See [Clustering](/docs/gridgain/gridgain-8/architecture/cluster/clustering.md).

**Native persistence** GridGain's built-in disk store that keeps a superset of the data on disk so the cluster survives restarts and crashes. See [Native Persistence](/docs/gridgain/gridgain-8/architecture/storage/native-persistence.md).

**Partition** A subset of a cache's data. Partitions are distributed across server nodes to balance storage and load. See [Data Partitioning](/docs/gridgain/gridgain-8/architecture/data-distribution/data-partitioning.md).

**Rebalancing** The redistribution of partitions across nodes when the cluster topology changes. See [Data Rebalancing](/docs/gridgain/gridgain-8/architecture/replication-and-consistency/data-rebalancing.md).

**Server node** A node that stores data and performs computations; the base storage and compute unit of a cluster.

**Thick client** A client that joins the cluster via its internal protocol, is aware of data distribution, and supports the full GridGain API.

**Thin client** A lightweight client that connects over a binary protocol with a limited but broad API and no dependency on the JVM.

**WAL (write-ahead log)** The log GridGain writes changes to before applying them, used to recover a consistent state after a crash when native persistence is enabled.