Abstract:To bridge the increasing processor-disk performance gap, buffer caches are used in both storage clients (e.g. database systems) and storage servers to reduce the number of slow disk accesses. These buffer caches need to be managed effectively to deliver the performance commensurate to the aggregate buffer cache size. To address this problem, two paradigms have been proposed recently to collaboratively manage these buffer caches together: the hierarchy-aware caching maintains the same I/O interface and is fully transparent to the storage client software, and the aggressively-collaborative caching trades off transparency for performance and requires changes to both the interface and the storage client software. Before storage industry starts to implement collaborative caching in real systems, it is crucial to find out whether sacrificing transparency is really worthwhile, i.e., how much can we gain by using the aggressively-collaborative caching instead of the hierarchy-aware caching? To accurately answer this question, it is required to consider all possible combinations of recently proposed local replacement algorithms and optimization techniques in both collaboration paradigms.Our study provides an empirical evaluation to address the above questions. Particularly, we have compared three aggressively-collaborative approaches with two hierarchy-aware approaches for four different types of database/file I/O workloads using traces collected from real commercial systems such as IBM DB2. More importantly, we separate the effects of collaborative caching from local replacement algorithms and optimizations, and uniformly apply several recently proposed local replacement algorithms and optimizations to all five collaboration approaches.When appropriate local optimizations and replacement algorithms are uniformly applied to both hierarchy-aware and aggressively-collaborative caching, the results indicate that hierarchy-aware caching can deliver similar performance as aggressively-collaborative caching. The results show that the aggressively-collaborative caching only provides less than 2.5% performance improvement on average in simulation and 1.0% in real system experiments over the hierarchy-aware caching for most workloads and cache configurations. Our sensitivity study indicates that the performance gain of aggressively-collaborative caching is also very small for various storage networks and different cache configurations. Therefore, considering its simplicity and generality, hierarchy-aware caching is more feasible than aggressively-collaborative caching.

Study of database buffer management approaches: Towards the development of a data mining based strategy

Managing Multiuser Database Buffers Using Data Mining Techniques

DEAM：Decoupled, Expressive, Area-Efficient Metadata Cache

What can Emerging Hardware do for your DBMS Buffer?

From Cooperative Scans to Predictive Buffer Management

Key Techniques in Design of DBMS Buffer Manager

“Anti-Caching”-based elastic memory management for Big Data

Multi-Tier Buffer Management and Storage System Design for Non-Volatile Memory

Buffer Allocation Algorithms for Embedded Real-Time Streaming File System

In-Memory Big Data Management and Processing: A Survey

Persistent Buffer Management with Optimistic Consistency

Empirical Evaluation of Multi-Level Buffer Cache Collaboration for Storage Systems

Real-time performance analysis of network buffer under multi-core scheduling platform

L2BM: Switch Buffer Management for Hybrid Traffic in Data Center Networks

HFB: A page-block hybrid buffer manager for flash memory

Caching With Finite Buffer and Request Delay Information: A Markov Decision Process Approach

A Cost-aware Buffer Management Policy for Flash-based Storage Devices

PWLRU: a Buffer Replacement Algorithm for Flash-Based Database

Receiving Buffer Adaptation for High-Speed Data Transfer

Energy-Aware Disk Storage Management: Online Approach with Application in DBMS

H/S co-design of embedded DBMS with high speed buffer support