A simple analysis of the LRU buffer policy and its relationship to buffer warm-up transient

Bhide, Anupam; Dan, Asit; Dias, Daniel M.

doi:10.1109/icde.1993.344070

Cited by 18 publications

(29 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Therefore, it is important for the proposed models to capture the number of disk page accesses when a (typically, LRU) buffer is available. The behavior of LRU buffers in general database environments has been analyzed in Bhide et al [1993]. Subsequent work [Huang et al 1997;Leutenegger and Lopez 2000] that addresses the buffered performance of specific index structures is based on Bhide et al's [1993] findings that, after the buffer warmup period (i.e., all buffer pages have been loaded), the probability of locating a page in the buffer does not vary significantly.…”

Section: Introducing a Lru Buffermentioning

confidence: 99%

“…According to Bhide et al [1993], the probability of (2) approximates the probability that the node is not accessed by any of the n q queries during the buffer warmup period. Since this probability is [1 − prob(s i , q)] n q for a level i node, the probability for the node to be loaded from the disk (to answer query q) is given by prob(s i , q) · (1 − prob(s i , q)) n q .…”

Section: Introducing a Lru Buffermentioning

confidence: 99%

See 1 more Smart Citation

Cost models for overlapping and multiversion structures

Tao

Papadias

Zhang

2002

ACM Trans. Database Syst.

View full text Add to dashboard Cite

Overlapping and multiversion techniques are two popular frameworks that transform an ephemeral index into a multiple logical-tree structure in order to support versioning databases. Although both frameworks have produced numerous efficient indexing methods, their performance analysis is rather limited; as a result there is no clear understanding about the behavior of the alternative structures and the choice of the best one, given the data and query characteristics. Furthermore, query optimization based on these methods is currently impossible. These are serious problems due to the incorporation of overlapping and multiversion techniques in several traditional (e.g., financial) and emerging (e.g., spatiotemporal) applications. In this article, we reduce performance analysis of overlapping and multiversion structures to that of the corresponding ephemeral structures, thus simplifying the problem significantly. This reduction leads to accurate cost models that predict the sizes of the trees, the node/page accesses, and selectivity of queries. Furthermore, the models offer significant insight into the behavior of the structures and provide guidelines about the selection of the most appropriate method in practice. Extensive experimentation proves that the proposed models yield errors below 5 and 15% for uniform and nonuniform data, respectively.

show abstract

Section: Introducing a Lru Buffermentioning

confidence: 99%

Section: Introducing a Lru Buffermentioning

confidence: 99%

Cost models for overlapping and multiversion structures

Tao

Papadias

Zhang

2002

ACM Trans. Database Syst.

View full text Add to dashboard Cite

show abstract

“…Moreover, our analysis provides a cost model of a typical (non-uniform) workload using the so-called biased query model [22], which assumes that queries are more probable in high-density areas of the address space. Moreover, we have successfully extended our cost model to other distance join queries as buffer queries [9] and to the study of including buffering (LRU buffer model, [6]) in K-CPQ using R-trees, which theoretical results have been very similar to the experimental ones in terms of buffer hit probability and number of R-tree node accesses.…”

Section: Introductionmentioning

confidence: 74%

“…3a for d = 2. Note the likeness of V M 1 ÈM 2 with the expression of join index selectivity, see formula (6).…”

Section: Number Of Node Accesses For K-cpqmentioning

confidence: 99%

See 1 more Smart Citation

Cost models for distance joins queries using R-trees

Corral

Manolopoulos

Theodoridis

et al. 2006

Data & Knowledge Engineering

View full text Add to dashboard Cite

The K-Closest-Pairs Query (K-CPQ), a type of distance join in spatial databases, discovers the K pairs of objects formed from two different datasets with the K smallest distances. Recently, branch-and-bound algorithms based on R-trees have been developed in order to answer K-CPQs efficiently. For query optimization purposes, analytical models are needed to estimate the processing cost of a specific query in order to evaluate alternative execution plans. In this paper, we combine techniques that have been used for the analysis of nearest neighbor and spatial join queries, and derive the performance cost (in terms of disk accesses) of K-CPQs using R-trees. Moreover, we present two interesting extensions of the cost model for K-CPQs, one exploiting the buffering management using R-trees and another for a second type of distance join, the socalled buffer queries. The proposed cost models are verified under a variety of distributions in 2-dimensional space on both synthetic and real datasets, shown to achieve accurate estimations of the measured experimental results.

show abstract

An Analytic Study of Caching in Computer Systems

Buck

Singhal

1996

Journal of Parallel and Distributed Computing

View full text Add to dashboard Cite

A simple analysis of the LRU buffer policy and its relationship to buffer warm-up transient

Cited by 18 publications

References 8 publications

Cost models for overlapping and multiversion structures

Cost models for overlapping and multiversion structures

Cost models for distance joins queries using R-trees

An Analytic Study of Caching in Computer Systems

Contact Info

Product

Resources

About