CMSC 818S
DataCutter Project
Proposal due Thursday, March 18, 6:00PM
Interim Report due Tuesday, April 13, 6:00 PM
Class Demos, starting May 6
Final Report due Tuesday, May 11, 6:00 PM

Proposal

You must write a short proposal (1-2 pages), describing the application you will implement using the DataCutter middleware.  The application you choose should be one with functionality that can be partitioned into a set of interacting components/filters, with data flowing between the components on DataCutter streams.  You will also need a reasonable sized dataset to test your application on.  Since you have a 500MB disk quota on the cluster (redleader), your dataset should be smaller than that.  Example applications that have previously been implemented with DataCutter include the Virtual Microscope, satellite data analysis, image similarity matching, census data analysis, photo mosaicking, and weather trend analysis.

In addition to the data server you will implement using DataCutter, you will also have to provide a GUI client for generating requests/queries to the server, and viewing results.  You can build this however you want, but keep it relatively simple and use good existing tools (e.g., Java Swing, or other packages for building GUIs).

Interim Report

For the interim report, you should expand on your proposal to make explicit the details of your DataCutter application.  That includes a precise description of the data you are using (and how much data you have), and of the queries you will support.  You should also fully describe the DataCutter filters you are implementing to process the queries.  Some description of the user interface for the application is also required.  The report should expand on the proposal, so will be longer, perhaps 3 or 4 pages.

Runtime environment

You will run your DataCutter application on the red/blue cluster, so that's where you should do your development work.  DataCutter is installed on redleader in ~chansen/work/818-DataCutter/ , with documentation in doc/ , binaries in bin/ , etc.  You should be able to set up your environment to use the software directly from that directory, but use your filter code and configuration files from your directory.  See the documentation for details.  Questions about DataCutter details should be directed to Christian Hansen.

Overall project

  1. You are working in pairs on the project, so when submitting your proposal and interim report  make sure it's clear who your partner is.
  2. The final part of the project is a report (5 pages max) describing the design and implementation of your project.  There will also be live demonstrations of the projects in class, near the end of the semester (starting May 6).
  3. DataCutter documentation and API specification are available online here.