Maintaining spatial data sets in distributed-memory machines

Abstract

Proposes a distributed data structure for maintaining spatial data sets on message-passing, distributed memory machines. The data structure is based on orthogonal bisection trees and it captures relevant characteristics of parallel machines. The operations we consider include insertion, deletion and range queries. We introduce parameters to control how much imbalance is tolerated at each processor and to specify the load to be achieved during balancing. When balancing, we first broadcast point counts of a data-dependent partition of the data. Based on this partition, we propose load balancing methods with different communication and computation requirements. We present initial experimental results for the Cray T3D.