Tracking semantic relationships for effective data management in home networks

  • Authors:

  • Ashok Anand;Aaron Gember;Aditya Akella;Vyas Sekar

  • Affiliations:

  • University of Wisconsin, Madison, Madison, WI, USA;University of Wisconsin, Madison, Madison, WI, USA;University of Wisconsin, Madison, Madison, WI, USA;Carnegie Mellon University, Pittsburgh, PA, USA

  • Venue:
  • Proceedings of the 2010 ACM SIGCOMM workshop on Home networks
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

The amount of data that home users generate, store, and peruse has grown significantly in the past few years. Increasingly, organizing this huge amount of data - in order to make it easy to browse, query and access - is becoming challenging. Many recent proposals have emphasized the importance of data management in home networks and proposed mechanisms for managing replicas across devices to increase availability. Essentially, they capture the relationship "is copy of" between files across devices. However, files can be semantically related. Users are often interested in finding data that has such semantic relationships; tracking these relationships helps users to effectively search based on content or human-understandable context, organize data and manage the limited storage while ensuring availability of information. However, inferring semantic relationships just based on user-defined tags and file names can be challenging, since users may not follow any standard or unique naming conventions. We argue that such semantic relationships should be derived on the basis of content itself, and propose to leverage recent developments in multimedia processing literature, with minimal user involvement. The decentralized, heterogeneous and dynamic operational environment of home networks present interesting systems and network challenges. In this paper, we have highlighted several candidate designs and system-optimizations that can help build an effective semantic-aware data management for home networks. As ongoing work, we are working on a prototype implementation of a decentralized data management system.