An Introduction to the IVC Software Framework

1
(No Transcript)
2
An Introduction to the IVC Software Framework
Bruce Herr (Lead Developer) Shashikant Penumarthy
(Team Lead)
November 2nd, 2005
3
Database
Software

• Diverse datasets

• Web Interface
• Search
• Analysis
• Download

Learning Modules
Compute Resources
4
Database
Software

• Diverse datasets

• Data analysis
• Modeling
• Visualization
• Preprocessing
• Tools

• Web Interface
• Search
• Analysis
• Download

Learning Modules
Compute Resources
5
Database
Software

• Diverse datasets
• Web Interface
• Search
• Analysis
• Download

• Data analysis
• Modeling
• Visualization
• Preprocessing
• Tools

Learning Modules
Compute Resources

• Computing access
• 2 Sun V1280s with 12x900Mhz CPUs and 96GB RAM
  each
• 1 Sun V880 with 4 CPUs and 8GB RAM
• 2 Sun V490s with 4x1.35Ghz CPUs and 16GB RAM
  each
• Etc

6
Database
Software

• Diverse datasets
• Web Interface
• Search
• Analysis
• Download

• Data analysis
• Modeling
• Visualization
• Preprocessing
• Tools

Compute Resources
Learning Modules

• Computing access

• Learn about InfoVis
• How to use algorithms
• What algorithms to use
• Application of InfoVis

7
Database
Software

• Diverse datasets
• Web Interface
• Search
• Analysis
• Download

• Data analysis
• Modeling
• Visualization
• Preprocessing
• Tools

Learning Modules
Compute Resources

• Computing access

• Learn about InfoVis
• How to use algorithms
• What algorithms to use
• Application of InfoVis

8
Motivation for the IVC Software Framework
9

• The Problem
• Lots of diverse tools
• Diverse data formats
• Algorithms rewritten ad infinitum
• Bad/unintuitive user interfaces
• Scientists unwilling to share code

10

• The Solution
• An open-source tool for integration of
  algorithms
• Independence
• Platform independence
• GUI toolkit independence
• Data model independence
• Incentives for scientists to contribute and use
  it

11
IVC Software Framework Architecture
12

• IVC Software Framework Goals
• Ease of use
• Ease of development
• Scalability
• Extensibility

13
IVC Software Framework Features

• Ease of Use
• Graphical installer/uninstaller
• GUI based interface
• Consistent GUI for algorithms
• Plug-ins (data, algorithms, viz, etc..) are
  added through the GUI
• All updating done through the GUI
• Session logging

14
IVC Software Framework Features

• Ease of Development
• Graphical code-generating templates
• A simple and well defined Java API
• Support for non-Java based algorithms
• Plug-in framework based off of Eclipse RCP
  platform
• Well-documented, supported, and established
  platform
• Multi-platform compatibility

15
IVC Software Framework Features

• Scalability
• Plug-in system is based on Eclipse RCP
• Built for handling large amounts of plug-ins
• Most scalability concerns are at the plug-in
  level
• Can conserve memory by not fully loading
  datasets into memory

16
IVC Software Framework Features

• Extensibility
• Everything is a plug-in
• API is a plug-in
• GUI is a plug-in and is swappable
• Algorithms, visualizations, etc are plug-ins
• Data persister plug-ins support new data formats
• Datasets are plug-ins
• Extend features through the plugin.xml schema

17
IVC Software Framework Features

• Future Extensions
• Scripting support
• Branding support
• Visualization framework
• Interfacing with existing tools (eg. R)
• Replayable logs
• Meta-data
• Different user-interfaces
• Command line
• Text/console
• Web-based