Understanding query quality in dynamic networks


Understanding query quality in dynamic networks

Show full record

Title: Understanding query quality in dynamic networks
Author: Rajamani, Vasanth
Abstract: With the proliferation of laptops, smart phones, sensors and other small devices, our physical environment is increasingly networked. Applications in a variety of problem domains (e.g., intelligent construction, traffic monitoring, smart homes, etc.) need to efficiently and seamlessly execute on top of such emerging infrastructure. Such infrastructure tends to be unreliable, and the network configuration changes constantly (network hosts depart and reemerge frequently). Consequently, software has to be able to react to these changes continuously and change its behaviors accordingly. In this dissertation, I introduce PAQ (Persistent Adaptive Query), a middleware designed to ease the programming burden associated with writing such applications. PAQ employs a novel style of query-driven application development that allows programmers to build pervasive applications by employing persistent queries--queries that continuously monitor the environment. The dissertation discusses the design and implementation of a new middleware model that allows programmers to write high level specifications abstracting away several tedious implementation details. PAQ employs both novel protocols that automatically tag the quality of information obtained from the network and statistical techniques to post-process and smooth the data. The goal of this research is to ease the software engineering challenges encountered during the construction and deployment of several applications in emerging pervasive computing environments thorough the use of a query-driven application development paradigm.
Department: Electrical and Computer Engineering
Subject: Query quality Adaptive query processing Dynamic networks Sensor networks Mobile adhoc networks Persistent Adaptive Query PAQ
URI: http://hdl.handle.net/2152/ETD-UT-2010-08-1745
Date: 2010-08

Files in this work

Size: 23.08Mb
Format: application/pdf

This work appears in the following Collection(s)

Show full record

Advanced Search


My Account