Identifier

etd-02132004-164149

Degree

Doctor of Philosophy (PhD)

Department

Computer Science

Document Type

Dissertation

Abstract

Grid computing is described as dependable, seamless, pervasive access to resources and services, whereas mobile computing allows the movement of people from place to place while staying connected to resources at each location. Mobile grid computing is a new computing paradigm, which joins these two technologies by enabling access to the collection of resources within a user's virtual organization while still maintaining the freedom of mobile computing through a service paradigm. A major problem in virtual organization is needs mismatch, in which one resources requests a service from another resources it is unable to fulfill, since virtual organizations are necessarily heterogeneous collections of resources. In this dissertation we propose a solution to the needs mismatch problem in the case of high energy physics data. Specifically, we propose a Quadtree Dictionary (QTD) algorithm to provide lossless, multi-resolution compression of datasets and enable their visualization on devices of all capabilities. As a prototype application, we extend the Integrated Spectral Analysis Workbench (ISAW) developed at the Intense Pulsed Neutron Source Division of the Argonne National Laboratory into a mobile Grid application, Mobile ISAW. In this dissertation we compare our QTD algorithm with several existing compression techniques on ISAW's Single-Crystal Diffractometer (SCD) datasets. We then extend our QTD algorithm to a distributed setting and examine its effectiveness on the next generation of SCD datasets. In both a serial and distributed setting, our QTD algorithm performs no worse than existing techniques such as the square wavelet transform in terms of energy conservation, while providing the worst-case savings of 8:1.

Date

2004

Document Availability at the Time of Submission

Release the entire work immediately for access worldwide.

Committee Chair

Aiichiro Nakano

DOI

10.31390/gradschool_dissertations.2381

Download

Share

COinS