Understanding Application-Level Caching in Web Applications: A Comprehensive Introduction and Survey of State-of-the-Art

TL;DR

This paper provides a comprehensive survey of application-level caching in web applications, covering design, implementation, management, and adaptive strategies to enhance performance and scalability.

Contribution

It offers an extensive overview of state-of-the-art approaches, including static and adaptive caching solutions, serving as a guide for researchers and developers.

Findings

Survey of existing application-level caching approaches

Analysis of static and adaptive caching strategies

Guidance for designing and managing caching solutions

Abstract

A new form of caching, namely application-level caching, has been recently employed in web applications to improve their performance and increase scalability. It consists of the insertion of caching logic into the application base code to temporarily store processed content in memory, and then decrease the response time of web requests by reusing this content. However, caching at this level demands knowledge of the domain and application specificities to achieve caching benefits, given that this information supports decisions such as what and when to cache content. Developers thus must manually manage the cache, possibly with the help of existing libraries and frameworks. Given the increasing popularity of application-level caching, we thus provide a survey of approaches proposed in this context. We provide a comprehensive introduction to web caching and application-level caching, and present state-of-the-art work on designing, implementing and managing application-level caching. Our focus is not only on static solutions but also approaches that adaptively adjust caching solutions to avoid the gradual performance decay that caching can suffer over time. This survey can be used as a start point for researchers and developers, who aim to improve application-level caching or need guidance in designing application-level caching solutions, possibly with humans out-of-the-loop.