Providing caching abstractions for web applications
Author(s)Gupta, Priya, S.M. Massachusetts Institute of Technology
Massachusetts Institute of Technology. Dept. of Electrical Engineering and Computer Science.
Nickolai Zeldovich and Samuel R. Madden.
MetadataShow full item record
Web-based applications are used by millions of users daily, and as a result a key challenge facing web application designers is scaling their applications to handle this load. A crucial component of this challenge is scaling the data storage layer, especially for the newer class of social networking applications that have huge amounts of shared data. Caching is an important scaling technique and is a critical part of the storage layer for such high-traffic web applications. Usually, building caching mechanisms involves significant effort from the application developer to maintain and invalidate data in the cache. In this work we present CacheGenie, a system which aims to make it easy for web application developers to build caching mechanisms in their applications. It achieves this by proposing high-level caching abstractions for frequently observed query patterns in web applications. These abstractions take the form of declarative query objects, and once the developer defines them, she does not have to worry about managing the cache (i.e., insertion and deletion) or maintaining consistency (e.g., invalidation or updates) when writing application code. We designed and implemented CacheGenie in the popular Django web application framework, with PostgreSQL as the database backend and memcached as the caching layer. We use triggers inside the database to automatically invalidate or keep the cache synchronized, as desired by the developer. We have not made any modifications to PostgreSQL or memcached. To evaluate our prototype, we ported several Pinax web applications to use our caching abstractions and performed several experiments. Our results show that it takes little effort for application developers to use CacheGenie, and that caching provides a throughput improvement by a factor of 2-2.5 for read-mostly workloads.
Thesis (S.M.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2010.Cataloged from PDF version of thesis.Includes bibliographical references (p. 99-101).
DepartmentMassachusetts Institute of Technology. Dept. of Electrical Engineering and Computer Science.
Massachusetts Institute of Technology
Electrical Engineering and Computer Science.