Three compute techniques that run on the node that owns the key: broadcast with local scan, single-key affinityCall, and cross-cache affinityCall. A scale experiment from 1,000 to 100,000 customers prints wall-clock and byte counts for each approach against pull-to-client.
Three mechanisms control where related entries live in the cluster: composite keys with @AffinityKeyMapped, the AffinityKey wrapper, and CacheKeyConfiguration. Run all three against a Customer and Invoice schema, watch each produce the same colocation outcome, and learn which to pick for a real schema.
Pick a partition count for a partitioned cache on Apache Ignite 2 or GridGain 8. Default is 1024. This guide covers when to accept it, when to size up, and the formula that maps cache row count to a safe partition number.
Diagnose and fix a hot partition on Apache Ignite 2 or GridGain 8. Three-step procedure identifies the partition, its owner, and the skew pattern. Matched fixes cover salting, affinity-field changes, cache-level overrides, schema redesign, and replication.
Watch a cache spread 10,000 entries across three nodes and then four. Use the Affinity API to see which node owns each key, learn the vocabulary of partitions and primaries, and establish the mental model every cache user needs.
Colocation fails silently. Three verification techniques prove a running cache is wired correctly before a bug reaches production: the Affinity runtime API, SYS.CACHE_GROUPS and SYS.PARTITION_STATES system views, and a per-node local scan that catches orphan rows.