综合论文：面向列存储vs.面向列存储（英文）.pdfVIP

Column-Stores vs. Row-Stores: How Different Are They Really? Daniel J. Abadi Samuel R. Madden Nabil Hachem Yale University MIT AvantGarde Consulting, LLC New Haven, CT, USA Cambridge, MA, USA Shrewsbury, MA, USA dna@ madden@ nhachem@ ABSTRACT General Terms There has been a significant amount of excitement and recent work Experimentation, Performance, Measurement on column-oriented database systems (“column-stores”). These database systems have been shown to perform more than an or- Keywords der of magnitude better than traditional row-oriented database sys- tems (“row-stores”) on analytical workloads such as those found in C-Store, column-store, column-oriented DBMS, invisible join, com- data warehouses, decision support, and business intelligence appli- pression, tuple reconstruction, tuple materialization. cations. The elevator pitch behind this performance difference is straightforward: column-stores are more I/O efficient for read-only 1. INTRODUCTION queries since they only have to read from disk (or from memory) Recent years have seen the introduction of a number of column- those attributes accessed by a query. oriented database systems, including MonetDB [9, 10] and C-Store [22]. This simplistic view leads to the assumption