I ran a geospatial query on Apache Ignite successfully. But I didn't understand how their partitioning work. How does Apache Ignite partition the spatial data among nodes when we use PARTITIONED CacheMode? Do they use any partitioning technique like Grid or Quad-tree? I saw that they are creating 1024 partitions for each data set. How can I change the number of partitions? I already read their documents but I didn't find anything about this. Any suggestion or document links will be appreciated.
0
votes
1 Answers
0
votes
Partitioning is performed on key basis using Rendezvous hashing. Apache Ignite is key-value based.
You can change its properties by specifying affinityFunction in CacheConfiguration.
Partitioning is usually wholly unrelated to spatial indexing since spatial index is secondary.
