Memory-Contention Responsive Hash Joins

In order to maximize systemperformancein environments with fluctuating memory contention, memory-intensive algorithms such as hash join must gracefully adapt to variations in available memory. Mixed workloads, creating fluctuations of erratic frequencyand magnitude,makeresponsivenessto memorycontentionparticularly important. Previous studies on adaptable hash joins have focused on lowering I/O costs by reducing the I/O volume, as measuredin the number of pages,by spilling partitions from memoryto disk and then restoring them into memory if more memorybecomesavailable. In this paper,wepresentmemory-contentionresponsivehashjoins that (i) reducethe amountof time spenton I/O by using large I/O buffers, or clusters, (ii) dynamically vary the cluster sizein responseto fluctuations in memory availability, and (iii) employ earlier techniquesof dynamic destagingandrestoration. Our simulation results demonstratethat these combined techniquesprovide better performancethan previous algorithms, pa...

• Type: Conference
• Year: 1994
• Where: VLDB
• Authors: Diane L. Davison, Goetz Graefe
• Keywords: Database, Hash Joins, Large I/o Buffers, Memory Contention, VLDB 1994