Concluded Projects

:: SPACE:Secure Protocol for Address-Book based Connection Establishment  

The past decade has seen a proliferation of consumer devices such as mobile phones with integrated wireless technologies such as Bluetooth and Wi-Fi. While this trend has suggested the possibility of peer-to-peer mobile applications (e.g., collaboration, gaming), this potential has remained large unrealized thus far. We identify two reasons for this: first, existing connection establishment protocols for Bluetooth and IEEE 802.11 have known security flaws, and can be compromised using well-known techniques and off-the-shelf hardware. Second, these protocols require explicit user intervention, like entering a passkey, which is often cumbersome.

SPACE is an application-level protocol for secure automatic ad-hoc connection-establishment between two devices based on their address book entries. Our protocol is based on the simple premise that if two people have each other’s contact details in their address books, they probably know and trust each other in some limited way, and this can form a basis for a trust relationship between their devices, without additional user intervention. Our protocol is resistant to specific security attacks and can accommodate privacy concerns. A position paper on SPACE appeared at the ACM HotNets 2006 workshop.

Project Members

  • Ganesh Ananthanarayanan
  • Ramarathnam Venkatesan
  • Prasad Naldurg
  • Sean Blagsvedt
  • Adithya Hemakumar
  • Karthik Kambatla


:: WiFiAds:A Novel Approach for Delivering Location-Sensitive Advertisements Using Wi-Fi Networks

The tremendous growth of Wi-Fi networks in recent years provides a unique opportunity for delivering location-specific information such as advertisements to users. Existing approaches to doing so depend on first determining a user’s location and conveying this information to a backend service for the appropriate ads to then be delivered to the user. However, these dependencies can be problematic: it may be difficult to locate the user automatically and with sufficient accuracy, and the user may not have Internet connectivity in many settings where they are nevertheless interested in receiving ads (e.g., while driving through a shopping district).

We propose a new approach for distributing location-sensitive advertisements to Wi-Fi devices, which relies on a clever use of certain features of the Wi-Fi protocol to "push" out ads to clients in the vicinity of a network. Our approach offers three main advantages over existing approaches. First, it does not require the client device to reveal any information in order to receive location-sensitive ads. Second, it does not require the client to have Internet connectivity. In fact, we can deliver ads even when the client is connected to the Internet via a competitor’s Wi-Fi network. Third, it allows advertisers to supply dynamic information to consumers in real-time; for example, a popular restaurant can continuously advertise the expected wait time to all wireless clients in its vicinity. A position paper on the technology underlying WiFiAds appeared at the IEEE HotMobile 2007 workshop.

Project Members

MSR India:

  • Lenin Ravindranath

MSR Redmond:

  • Ranveer Chandra
  • Jitu Padhye
  • Alec Wolman

:: Senslide: A Sensor Network-Based Landslide Prediction System

Senslide is aimed at predicting rockslides in the areas like the Konkan and the Himalayas. The project had its genesis in IIT Bombay when researchers in Earth Sciences and Electrical Engineering suggested the use of a wireless sensor field consisting of cheap sensors based on strain gauges to predict rockslides. Apart from interesting aspects in understanding rock behaviour, and in signal processing, there are significant problems to overcome in providing fault-tolerance, low power consumption, and reasonable accuracy of prediction in such a distributed system. Our initial work in this area is promising and is summarized in a poster abstract that appeared at the ACM SenSys 2005 conference.

Project Members

  • Anmol Sheth (Univ. of Colorado Boulder, USA)
  • Chandu Thekkath

External Collaborators

  • Uday Desai (IIT Bombay)
  • T N Singh (IIT Bombay)

:: OWeb: A Framework for Offline Web Browsing

Internet browsing has traditionally been dependent on the real-time network availability and speed. This dependency becomes a significant constraint when accessing the web over slow and intermittent networks. Accessing large chunks of data, or streaming content across the web often incurs problems with latency, interruptions, and poor quality, at least where network connectivity is questionable. In this project, we explore ways in which the Really Simple Syndication (RSS) feeds could be utilized for improving the Internet browsing experience.

Successful models for offline data access include SMS and E-mail synchronization over mobile phones. The important characteristic of this class of applications is that the end-user never needs to wait for access to data. The content is always available and offline, which results in a positive user experience, since the user is insulated from the vagaries of the network. To this end, our OWeb system incorporates a number of techniques including intelligent prefetching, graceful handling of intermittent network connectivity, and template identification and page stitching based on RSS feeds. We also propose extensions to the RSS protocol to enable a more effective utilization of the feeds by clients with intermittent connectivity. Our initial results are reported in an LA-WEB 2006 paper.

Project Members

  • Ganesh Ananthanarayanan
  • Sean Blagsvedt
  • Kentaro Toyama