Social Bookmarking Over Nosql. We Need Speed

Abstract

Becoming the backbone of some of the web’s biggest social networks, and few others systems holding a large amount of data, the NoSql concept stands its ground as the only alternative to storing data beside the SQL approach. Big web companies like Google, Amazon and Facebook developed non-relational databases that sacrifice consistency for availability, scalability and performance. We have developed a social bookmarking service based on NoSql concept, looking for performance, since expecting a large amount of data in our hands. A system providing its users with an easy way to organize, tag and share web content of interest. Additionally, generating recommendations and groups of similar web content, based on interests shown while using the bookmarking service. In this paper we concentrate on building a solid ground for performance comparison of two web applications, providing the same social bookmarking service, using a different concept to store its data. The expectations are that the NoSql based service will provide us with better performance that the SQL supported service. We concentrate on the potential performance benefit, as well as the possibility to gain additional advantage by being able to scale our data storage on multiple servers.