env 200616.1 envisioning information lecture 16 – distributed and collaborative visualization ken...

38
ENV 2006 16.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie [email protected]

Upload: evelyn-pereira

Post on 28-Mar-2015

215 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.1

Envisioning Information

Lecture 16 – Distributed and Collaborative Visualization

Ken [email protected]

Page 2: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.2

Outline of Lecture

• From Visualization to Computational Steering

• Distributed visualization– Extending dataflow across the network– Grid-based visualization and computational steering

• Collaborative visualization– Sharing the display screen– Sharing the visualization

Page 3: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.3

Dataflow Visualization Systems

• Visualization represented as pipeline:

– Read in data– Construct a visualization in

terms of geometry– Render geometry as image

• Realised as modular visualization environment

– IRIS Explorer is one example– Visual programming paradigm– Extensible – add your own

modules

http://www.nag.co.uk

data visualize render

Page 4: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.4

• IRIS Explorer is one of a family of similar visualization systems

• First product was AVS– Still major player but no longer visual programming– http://www.avs.com

• Amira, IBM Open Visualization Data Explorer (DX), IRIS Explorer– visual programming based : plug, play, throw away– application decomposed as set of modules, configured at run-time (blur

between building and running an application)– open : user can write modules– low-cost

Visualization Software Environments

Page 5: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.5

IBM Open Visualization Data Explorer – now OpenDX

• Released around 1991 by IBM

• Made open source in 1999– www.opendx.org

• A major use of it has been for weather visualization

http://www.research.ibm.com/weather/

Page 6: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.6

Amira

• More recent product

• Increasing use for medical applications..

• .. But also engineering including CFD

• Marketed by TGS

• www.tgs.com

Page 7: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.7

vtk - Visualization Toolkit

• vtk is a programming - based toolkit

• Open source C++ library

• www.kitware.com

Page 8: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.8

.. And there are many others

http://www.kdnuggets.com/software/visualization.html

Page 9: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.9

IRIS Explorer - Creating Your Own Modules

• It is possible to create your own modules • The mbuilder tool creates a wrapper around your own code• See:

http://www.nag.co.uk/visual/ie/iecbb/doc/html/unix-iemwg5-0.htm

Page 10: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.10

• IRIS Explorer can be driven either by GUI or by command line interface

• Commands can be grouped as a ‘script’ that IRIS Explorer runs– explorer -script <file>

• This allows Explorer to be run in batch mode, or to be driven by another application

• The scripting language is called Skm (pronounced as ‘scheme’)• Can be used interactively…• … in linux

– explorer -script %

• … in Windows, – use Skm editor (view menu)

Scripting - skm

Page 11: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.11

• To launch a module:(start “ReadImg”)

(start “DisplayImg”)

• To connect ports:(connect “ReadImg” “Output” “DisplayImg” “Input”)

• To start a map:(start-map “cfd”)

• See chapter 6 of User Manual on Web

http://www.nag.co.uk/visual/ie/iecbb/doc/html/unix-ug-chap06.htm

Creating a Simple Script

Page 12: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.12

Visualization and Simulation

• Visualization is a key tool in understanding the results of numerical simulations of complex physical phenomena

• Different modes of combining simulation and visualization:

– Post-processing

– Tracking

– Steering

Page 13: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.13

Linking Visualization and Simulation – Post Processing

• Post-processing

– Do the simulation and store results (step 1)

– Look at the results in a separate process (step 2)

– Revise the simulation (back to step 1)

simulation

data visualize render

Step 1

Step 2

PRO: study at your own pace

CON: must finish simulation first

Page 14: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.14

Linking Simulation and Visualization - Tracking

• Tracking

– Exploit extensibility of the dataflow visualization environment by including the simulation in the pipeline

– Track the behaviour of simulation as it runs

simulate visualize render

PRO: can abort fruitless simulations

Page 15: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.15

Linking Simulation and Visualization - Steering

• Computational steering:– By including a control module in the pipeline, we can direct the simulation in

response to the visualization

simulate visualize rendercontrol

PRO: not only can we track, we can alterthe actual course of the simulation

‘Human-in-the-loop’‘Human-in-the-loop’

Page 16: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.16

Computational Steering Environments

• Early visualization systems all have this extensibility feature and so can be used for steering

– IRIS Explorer for example

• New systems have emerged specifically to support steering

– SCIRun from Utah

Pressure profile for EHL contact

Page 17: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.17

Imagine this ….

• An explosion!

• A dangerous chemical escapes!

• Where is the fugitive pollutant headed?

• Who needs to be evacuated?

Page 18: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.18

Understanding What Will Happen

• Model the dispersion by solving system of PDEs

• Understand solution by visualization

• What if scenarios … need to be able to steer the simulation

• For example, what if the wind changes direction?

Page 19: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.19

Tracking the Pollution

Page 20: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.20

What can be Steered?

• Steering requires the writer of the simulation code to expose parameters that can be legitimately modified in the course of a run

– frequency of output of results– values of external influences that may vary over a simulation

• Not all parameters can be changed– time step used by numerical codes to achieve stability and/or accuracy

• Notion of backtracking is important in some simulations– Often you first observe, then wish to rewind a few timesteps, then replay with

different parameter settings

Page 21: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.21

Our Scenario

• We shall use this scenario to illustrate:

– Distributed visualization : we need to understand where the pollutant is headed in faster than real-time … therefore we need to run the simulation on a powerful compute resource

– Collaborative visualization : there is no time to collocate the scientist, the meteorologist, the politician or whoever needs to be involved … so we need to link people in over the network to allow them to visualize collaboratively

… while still using IRIS Explorer!

Page 22: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.22

Harnessing Remote Compute Resources – Grid Computing

Explorer on single host

Explorer on multiple hosts

Select remote host

Automatic authentication using: •Globus certificate

•SSH Key pair

Page 23: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.23

Simulation Runs Remotely

Page 24: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.24

A Tale

Page 25: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.25

The Monkey Gets the Nuts – Two Heads ARE Better than One

Thanks to Accra Academy, Ghana

Page 26: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.26

Collaborative Working

• Radical collocation has proved highly successful in a number of areas

– Space missions

– Safety critical software development

• Productivity doubled– Teasley et al, Univ of Michigan

• But this requires:– Social disruption

– Advance planning

– … and can end in tears

• Can we gain at least part of this success using electronic collaboration?

Page 27: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.27

Visualizing Collaboratively

• We need to move away from seeing collaborative visualization as a group around a display screen..

• .. Towards collaboration over a network

Collaborative visualization

Page 28: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.28

Collaborating in the Pollution Demonstrator

• Who needs to collaborate and in what way?

• Scientists and numerical modellers– Discuss amongst each other possible scenarios– Discuss need to pull in further Grid resources perhaps

• Meteorologist– Will play an active part in controlling the simulation

• Environmental agency decision makers– Need to analyse ‘what-if’ scenarios and construct presentations for

politicians

• Politicians, local authorities– Want to see clear presentation of consequences– Probably not interested in steering

Page 29: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.29

Sharing the Display Screen

• A very simple model is to broadcast the display screen of an application to a set of (passive) users

– Operating system level

– Screen image is broadcast using intelligent compression

– Only active user can enter input

data visualize render

internet

User A executes application

User B receives copy of user A desktop- does not execute application

data visualize render

Page 30: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.30

Sharing the Display Screen

• There are a number of available technologies for screen sharing

• VNC – Virtual Network Computing– Family of open source products evolved from original VNC development

by AT&T– RealVNC : www.realvnc.com (original development team)– tightVNC : www.tightvnc.com (new compression algorithms)– Heterogeneous

• Microsoft NetMeeting (and now MSN Messenger)

Page 31: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.31

Sharing Display Screen

• Advantages– Very simple concept – works for any application– Good for training– Good for presentation to a group

• Disadvantages– No independent working– Performance issues when rapid screen changes

• Variations– (1) Only one master – only one can control by mouse and keyboard input– (2) Any participant can input

Page 32: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.32

Sharing the Visualization

• This is a more flexible approach – and specific to dataflow visualization

• Each collaborator is an active participant in the visualization process

• Multiple, interlinked applications, where each collaborator runs their own application but data and parameter settings are programmed to be shared between the different applications

Page 33: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.33

internet

data visualize render

Sharing the Visualization

• Extends the dataflow model to interlink pipelines across the Internet

• Collaborative server provides the link

• So one user – for example - can send geometry to another person for viewing

collaborative server

share

share

render

Page 34: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.34

Programming the Collaboration

• It is useful to be able to program the collaboration

– To adapt to how people want to collaborate

– To adapt to network bandwidths

• Here raw data is exchanged so a different visualization can be created

internet

collaborative server

data visualize render

share

share

visualise render

Page 35: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.35

COVISA in action

sharing isosurface level

sharing data

Collaborator A Collaborator B

Page 36: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.36

Multiple, Interlinked Applications

• COVISA part of IRIS Explorer

• Advantages– Great flexibility– Independent working

• Disadvantages– Difficult to understand what the other user is doing

Page 37: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.37

Bring in the Meteorologist Remotely

Scientistin lab

Initiatecollaborativesession

Link inmeteorologistremotely

Page 38: ENV 200616.1 Envisioning Information Lecture 16 – Distributed and Collaborative Visualization Ken Brodlie kwb@comp.leeds.ac.uk

ENV 2006 16.38

Conclusions

• We have studied many aspects of scientific visualization:

– Applications and history– Different techniques for scalar

and vector data– Distributed and collaborative

visualization

• The practical work is giving experience in

– Exploratory visualization (what is going on?)

– Presentational visualization (here’s what is going on!)

• Finally, this afternoon, two case studies

– Exploration using parallel coordinates

– Focus and context for volume visualization