Term-weighting approaches in automatic text retrieval
Information Processing and Management: an International Journal
SOSP '91 Proceedings of the thirteenth ACM symposium on Operating systems principles
Computer Evaluation of Indexing and Text Processing
Journal of the ACM (JACM)
A vector space model for automatic indexing
Communications of the ACM
Practical File System Design with the Be File System
Practical File System Design with the Be File System
Optimal aggregation algorithms for middleware
Journal of Computer and System Sciences - Special issu on PODS 2001
Introduction to Information Retrieval
Introduction to Information Retrieval
Efficient B-tree based indexing for cloud data processing
Proceedings of the VLDB Endowment
| Hi-index | 0.00 |
Social network systems and services have become amazingly popular in recent years. This has resulted in huge amounts of data being published by users. At the same time, a great number of relationships between users, user groups, and (collections of) data items are constantly being established based on highly dynamic tagging activities by users. With this work we present the design and implementation of a special-purpose user-level file system, coined SNFS, designed to manage social-network entities (data items, users and their profiles, and tags) and their relationships. At the core of our approach lie tagging, indexing, and ranked retrieval (top-k) algorithms, allowing the key functionality to be provided in a timely manner. We discuss the core design and implementation features of SNFS and present a performance evaluation, exposing the key performance costs, and present alternative designs and implementations to overcome them. Finally, we provide a brief comparison with a well-known desktop search application, Beagle, and show, using real datasets, that for our envisaged queries SNFS provides significant performance gains.