Force Directed Layout of DNS heirarchy. Critical Infrastructure Protection
For the Workshop on "Information Visualization
Software Infrastructures" at IEEE 2004 Visualization,
Organized by Katy Börner, Indiana University, USA
and Jean-Daniel Fekete, INRIA, France
I.1) What functionality should a general InfoVis infrastructure
provide?
Information Visualization inherits many of it's requirements
from at least two precursor disciplines: Scientific Visualization and
Statistical Graphics with influences from Cartography, Graphic Design, et cetera.
There are a few significant differences introduced by the Information Visualization Problem:
I.2) What do you see as the main technical challenges for creating a central but flexible and universally useful (information) visualization software infrastructure (as opposed to 100 different ones)?
Please describe the (information) visualization software
infrastructure you are working on.
Los Alamos Naitonal Laboratory, in collaboration with University of New Mexico
and George Mason University have been building tool sets and frameworks with
the general intention of supporting information visualization (perceptualization)
solutions. Our frameworks and tools are generally intended for our
own use, but as we refine them, they become more suitable for general use.
We have some lessons learned in the technology as well as the general approach
to the problem of encoding abstract information for exploration and analysis.
Our work may be considered somewhat esoteric, but it does have grounding in several areas including computer science, cognitive and perceptual psychology, category theory, and linguistics/semiotics.
Immersion
We believe that a "sense of presence" and an identification
with the data being explored and analyzed is important. This can happen
through high-end VR technology but it can also happen through the proper use
of motivated representation and interaction on much lower-end hardware (like
a laptop).
Perceptualization
We believe that "Visualization" is the correct archetype for the larger, more encompassing
and useful concept of "Perceptualization". Not only is multi-sensory
perception (vision, sound, haptics, etc) implied, but also the lower perceptual
processing functions that occur below the linguistic/symbolic level.
Color, shape, motion, sound, etc... all have implications at a pre-linguistic
level which must be addressed. "The quick brown fox jumped over
the lazy dog" invokes a perceptual experience as well as a more semiotic/linguistic
experience.
Reification
This term means "to make the abstract real".
In this case, we refer to the act of binding perceptual/representational elements
(location, layout, color, shape, time, sounds, etc) to abstract data in a way
that gives it a "reality". We believe that doing this well involves
both consistent and motivated mappings between these two, and very likely with
mathematical transformations which yield more useful, meaningful, or consistent
quantities and qualities than the raw data. The use of "figurative expression",
particularly complex metaphors (or metaphor complexes) is the most obvious example,
but caricature and metonmy are other possibly familiar examples.
Our technical approach to this problem has been to build or adopt a
number of layers of software/frameworks.
At the bottom of our technological pyramid lies fairly low-level standards and
technologies such as C/OpenGL/Window managers/ MIDI, etc. Above
that is the Flatland visualization development framework which essentialy supports
loading and running multiple applications in a VR environment, inter application
communication, etc. On top of Flatland, we have a layer known as "Flux"
which supports a streaming data model and simple component architecture.
Built on top of Flux, is a representational abstraction for point, line and
surface glyphs, time and space layout/ordination, graph layout, etc. which is
also intended to include "reification" elements such as metaphor mapping, etc.
II.1) Project Name and Web Address
Flatland http://www.ahpcc.unm.edu/homunculus/
Flux: (no public information)
Reifier: (no public information)
II.2) Core Team Members
LANL: Decision Applications Visualization Team
Team Lead: Steve Smith; Flux
Architect: Paul Weber; Developers: David Hite, Bob
Gislason, Steve Linger, Chris Davis.
UNM Flatland: High Performance Computing Center
Visualization Group/Homonculous Project
Director: Dr. Thomas P. Caudell;
Staff; Panaiotis; Timothy B. Eyring, Chris Davis, Takeshi
Hakamata, Victor Vergera, Jim Holten, et alia
II.3) Project Start Date
Ongoing
II.4) Targeted User Group
Internal R&D
Decision Applications Developers
II.5) Supported User Tasks
Application by application details.
II.6) Major Features of the System Architecture
Multi-user, telecollaborative, immersive environment
with a component architecture for "Reification". We focus on huge, high
dimensional, heterogenous, complex, streaming data sets. We focus on problems
such as algorithm, graph, and network analysis.
II.7) Algorithms Provided
Data flow management, sorting, time-like variables, force
directed layout, high dimensional projection, grand tour, clustering...
II.8) Snapshot of the Interface
Force Directed Layout of DNS heirarchy.
Critical Infrastructure Protection
"Space Defense" metaphor for Network Intrusion Detection.
UI is primarily mediated through immersive devices, keyboards and
Head's Up display (not shown). Traditional GUIs have been
used but are contra-indicated in immersive spaces.
Screenshots of what are generally designed for "large field of view",
navigable,
dynamic and interactive metaphorical worlds do not present this work well.
II.9) Development Platform
Linux
II.10) Supported Operating Systems
Linux/Unix variants
Windows (Marginal)
MacOS X (marginal)
II.5) Software Dependencies/Required Libraries
OpenGL
SDL
G++
gmake, etc.
II.5) Current License
Flatland: LGPL
II.5) Number of Users/Downloads
Internal to team
II.5) Pros and Cons
Very rich environment for immersion,
interaction, representation.
High learning curve, not quite production ready.
II.5) Planned Work
Huge amounts of planned work including coupling with
Parallel R, Volume Visualization and GPU programming. Direction of work
depends heavily on funding/sponsors whom are fickle.
Please describe your main interest in participating in the workshop
We believe that the problem of Information Visualization
(or Visual Analytics or Abstract Perceptualization or Immersive Data Exploration
and Analysis) is very important to many disciplines and that early identification
of the central issues and a general layout of appropriate approaches is extremely
valuable. We have been working in a relative vacume and look forward to
being a more integral part of a larger community. This group appears to
have the potential of being the kernel of such a community.
Determining the feasibility of combining efforts to create one common, shared
IV infrastructure as opposed to 100s of underfunded or proprietary toolkits,
platforms and frameworks. Scouring for ideas for a common data protocol for
communication between plugins. Eliciting feedback about the IVC software architecture
with regard to extensibility and ensuring that it is future-proof.
We believe that there is a great deal of motivation for creating
a common, shared IV infrastructure, but our experience indicates that this is
only partially possible. We have some limited experience with trying
to create an environment similar to what is proposed here and are willing to
share our ideas, lessons learned, and in some cases code and algorithms.
We do not believe that what we have created is anything more than an R&D
prototype suitable for our own use and those of a motivated select group whose
requirements for immersion, real-time rendering and Reification are similar
to our own.
Send the completed paper by Sept. 30, 2004 to katy@indiana.edu and
Jean-Daniel.Fekete@inria.fr.