Memory Efficient Algorithm for Mining Recent Frequent Items in a Stream

2014 22nd Euromicro International Conference on Parallel, Distributed, and Network-Based Processing

2014

110

This paper proposes to use a frequency based cache admission policy in order to boost the effectiveness of caches subject to skewed access distributions. Given a newly accessed item and an eviction candidate from the cache, our scheme decides, based on the recent access history, whether it is worth admitting the new item into the cache at the expense of the eviction candidate.Realizing this concept is enabled through a novel approximate LFU structure called TinyLFU, which maintains an approximate representation of the access frequency of a large sample of recently accessed items. TinyLFU is very compact and light-weight as it builds upon Bloom filter theory.We study the properties of TinyLFU through simulations of both synthetic workloads as well as multiple real traces from several sources. These simulations demonstrate the performance boost obtained by enhancing various replacement policies with the TinyLFU eviction policy. Also, a new combined replacement and eviction policy scheme nicknamed W-TinyLFU is presented. W-TinyLFU is demonstrated to obtain equal or better hit-ratios than other state of the art replacement policies on these traces. It is the only scheme to obtain such good results on all traces.

Section: Approximate Counting Architecturesmentioning

confidence: 99%

TinyLFU: A Highly Efficient Cache Admission Policy

Einziger

2014 22nd Euromicro International Conference on Parallel, Distributed, and Network-Based Processing

2014

110

“…In [2,21], conservative updates were used in order to monitor network traffic and discover heavy hitters. The work of [18] uses conservative updates in order to detect recently active heavy hitters. In [9], conservative updates are used in order to approximate the LFU cache admission policy.…”

Section: Introductionmentioning

confidence: 99%

A formal analysis of conservative update based approximate counting

Einziger

2015 International Conference on Computing, Networking and Communications (ICNC)

2015

This paper presents a formal analysis of multiple popular approximate counting schemes that employ the conservative update policy, such as CU-Sketch and Minimal Increment Spectral Bloom Filters, under a unified framework. It is also shown that when applied to items picked from a skewed distribution, such as Zipf-like functions, the analysis follows very closely empirical results obtained through simulations. Furthermore, this paper's analysis is orders of magnitude more accurate than previously known analysis of approximate counting schemes.

TinyLFU

Einziger¹,

Manes³

2017

ACM Trans. Storage

This article proposes to use a frequency-based cache admission policy in order to boost the effectiveness of caches subject to skewed access distributions. Given a newly accessed item and an eviction candidate from the cache, our scheme decides, based on the recent access history, whether it is worth admitting the new item into the cache at the expense of the eviction candidate. This concept is enabled through a novel approximate LFU structure called TinyLFU , which maintains an approximate representation of the access frequency of a large sample of recently accessed items. TinyLFU is very compact and lightweight as it builds upon Bloom filter theory. We study the properties of TinyLFU through simulations of both synthetic workloads and multiple real traces from several sources. These simulations demonstrate the performance boost obtained by enhancing various replacement policies with the TinyLFU admission policy. Also, a new combined replacement and eviction policy scheme nicknamed W-TinyLFU is presented. W-TinyLFU is demonstrated to obtain equal or better hit ratios than other state-of-the-art replacement policies on these traces. It is the only scheme to obtain such good results on all traces.