Research
Currently I am working under Prof. Tarek Abdelzaher on developing distributed caching heuristics for sensor networks as a part of the NEST project.

 
Motivation
It will soon be possible to develop, economically, large numbers of small smart components that combine computing power, wireless communication capabilities, and specialized sensors and actuators. These components or nodes may be deployed in thousands to achieve a common mission. They may be used to monitor poorly accessible or dangerous environments such as the ocean floor, neighborhoods of volcanic activities, hostile territories (e.g., behind enemy lines) etc. In a typical scenario, an unmanned aircraft drops thousands of sensors behind enemy lines. These sensors have high failure rates and limited battery life. These sensors form groups and gather and propagate data about enemy vehicles, terrain information etc. Soldier have handheld devices which provide them with information by communicating with the sensor network through base stations.

The paradigm of a sensor network is vastly different from those of traditional internet-type networks running protocols like TCP/IP, UDP. The main bottlenecks are : battery power, limited communication bandwidth, low processing power, limited local memory. One important fall-out of these is that a sensor by itself cannot accomplish much, but as a group they can gather substantial information. The tradeoffs involved are optimizing the amount of traffic in the network, with in turn leads to decrease in battery power and hence network lifetime. There has been some amount of work in information gathering in sensor networks, saliently directed diffusion. Some work has been done on energy-aware routing. Our work focuses primarily on providing a service which shall try  minimize the traffic by placing the data at a near optimal position, at the same time making sure that invalidating or refreshing of data does not overload the network.

 

Objectives
We aim to develop distributed heuristics for caching/placement of data so that the communication is minimized. The main design issues here are :
  • Memory - The protocol should be lightweight, so as to operate within the memory constraints.
  • Distributed -  The computation should be distributed, because the nodes are severely limited in computational power.
  • Communication - The cached data should be placed such that the communication is minimized.
  • Consistency - The cached data should ideally provide the same consistency semantice as the original data.
  • Scalable - The performance should not degrade with increase in the popularity of the data.
  • Update mechanisms - The cached copies of the data should not be scattered about, so that trying to invalidate copies does not lead to flooding.
Publications
1. Data Placement for Energy Conservation in Wireless Sensor Networks - Sagnik Bhattacharya, Tarek Abdelzaher. Submitted to ICDCS 2002. (pdf)