Upload
ella-norman
View
215
Download
0
Embed Size (px)
Citation preview
Slide 1/20RCAC – Rosen Center For Advanced Computing
Real-time Data Delivery and Remote Visualization through Multi-layer
Interfaces
Vinaitheerthan Sundaram, Lan Zhao, Bedřich Beneš, Carol X. Song, Rakesh Veeramacheneni, Peter Kristof
Rosen Center For Advanced ComputingDepartment of Computer Graphics Technology
Purdue University
Work supported by:National Science Foundation
Slide 2/20RCAC – Rosen Center For Advanced Computing
NEXRAD II Data
• Next Generation Radar (NEXRAD) Level II Data (OR) Weather
Surveillance Radar (WSR-88D) Level II Data– This data contains a very fine temporal and spatial resolution of three
attributes: reflectivity, Doppler radial velocity and spectrum width
– These attributes are vital to understanding, monitoring and predicting severe
weather conditions
– There are 158 Radar Stations in the country
Acknowledgment: Figures are downloaded from websites
www.CCSU.edu and www.answers.com.
Slide 3/20RCAC – Rosen Center For Advanced Computing
NEXRAD II Data Generation• 3D structure in Radar Data
– Radars go through a programmed set of movements, which involve a continuous rotation over 360° in azimuth and a simultaneous increase in elevation by 1° to 3° per complete sweep
• Continuous NEXRAD Level II radar data stream– The radar data files vary in size from a few megabytes to tens of megabytes each,
depending on the weather conditions. The files are compressed using modified bzip2
– The temporal resolution is 4-5 minutes in severe weather vs. 9-10 minutes in calm weather
Slide 4/20RCAC – Rosen Center For Advanced Computing
Availability of NEXRAD II Data in Near Real-time
• NEXRAD II Data is available in real-time on the
TeraGrid through Purdue resource provider.
• Opportunity: – The real-time availability of high-resolution radar data provides
an exciting opportunity for a wide spectrum of users ranging
from basic ( students) to expert (researchers) if the radar data
can be accessed and visualized in 3D in a timely manner.
• However, catering to wide spectrum of users presents
unique challenges as the requirements for each user
differ.
Slide 5/20RCAC – Rosen Center For Advanced Computing
Talk Outline
• Motivation– Providing real-time data access and remote visualization for a
wide spectrum of users
• Challenges– A review of challenges in the state-of-the-art systems
• A Unique and Versatile System Design– Multi-layer Interfaces
– Multiple Service access points
• Back-end Architecture – The enabler– Parallel Data Pre-Processing and partial-volume caching
• Summary and Future Work
Slide 6/20RCAC – Rosen Center For Advanced Computing
Challenges
• Limitations in the State-of-the-art– Not handle large amounts of data from many stations over a
long time
– No direct interaction with the data for users
– Not accessible to general public because of complicated interfaces
– No access points to third party applications
• Challenges– Data management issues: Radar data streams at 50 MB/secs
– Native compression format of radar data
– Data processing: Computationally expensive processing
– Special-purpose hardware (GPU) required
Slide 7/20RCAC – Rosen Center For Advanced Computing
Different User Levels and their requirements
• Expert Users ( small group )– Perform in-depth investigation– Examples: Researchers, Emergency Management Personnel
• Learners/Casual Users ( large group )– Access and visualize data for educational or personal purposes– Examples: K12 and College Students, Public
• Advanced Users ( small group )– Explore and evaluate data but don’t have resources– Examples: Graduate Students, Researchers from other domains
• Users levels are NOT mutually exclusive– Expert User can be an advanced user when attending
conferences and can be a casual user when teaching a class
Slide 8/20RCAC – Rosen Center For Advanced Computing
System Design
Expert Users
Casual Users
Advanced Users
Slide 9/20RCAC – Rosen Center For Advanced Computing
LiveRadar3D Gadget• Web 2.0 technologies
– AJAX, Google Gadgets, Social Networking Applications
– Allows rapid dissemination of scientific tools to wide audience
• Live Radar 3D– Shows animated Flash movie of
3D visualization of the region near user’s zip
– Scalable because movies are pre-generated and stored at the webserver
– Granularity: 7 Regions (midwest, south, southwest .. )
Slide 10/20RCAC – Rosen Center For Advanced Computing
LiveRadar3D Desktop• Desktop client
– Written in C++– Runs on Linux / Windows – Can be run on standard GPU
cards– Uses pre-processed volumes– Leverages Teragrid processing
power and local GPU
• Advantages– Fast interactive 3D
manipulation– Scalable: Supports large number
of stations and large intervals of time
User selects Radar- stations and time period
• Usage Scenario
Tool connects to data-access interface and fetch
processed volumes
Tool renders on the local GPU
Slide 11/20RCAC – Rosen Center For Advanced Computing
LiveRadar3D VNC
• For users who – Don’t want to download a client
– Don’t have the resources such as GPU
• Uses VirtualGL/TurboVNC to enable remote 3D visualization
• A convenient way to do advanced interactive and collaborative visualization remotely
• Browser accessible• Similar to LiveRadar3D Desktop in functionality
– Allows full capability available to expert users
• Disadvantage – Needs server farm to scale
Slide 12/20RCAC – Rosen Center For Advanced Computing
3rd Party Applications Access
• Our architecture is modular and supports fine-grained service access points
• Enables developing interesting 3rd party applications such as– Weather prediction application can connect to data access
interface
– Custom 3D visualizations can be built on pre-processed volumes
Slide 13/20RCAC – Rosen Center For Advanced Computing
Services and Backend Data Architecture
• In our earlier work, we presented a system that – Accesses NEXRAD II data
– Processes it into render-able 3D volumes using Teragrid
– Visualizes using Texture-based volume rendering
• Disadvantages– On-demand processing => Slow for large amounts of data
– Single access point targeted at expert users
• Extensions:– Multiple services and access points
– Preprocessing data to improve response time and scalability
– Volume Caching for easy access and reuse
Slide 14/20RCAC – Rosen Center For Advanced Computing
Backend Data Flow Diagram
Slide 15/20RCAC – Rosen Center For Advanced Computing
Parallel Data Pre-Processing
• Partial 3D volumes – efficient data structure using Hash-maps
– spatial/temporal independence property => parallel generation
– can be quickly merged to form full 3D volumes that can be rendered
– two orders of magnitude smaller than actual data and much smaller than generating full 3D volumes
• Generation of partial volumes on Teragrid– Monitors the arrival of new data and pre-processes them and
stores the partial-volumes
– SRB archives the past-year partial volumes
Slide 16/20RCAC – Rosen Center For Advanced Computing
The fully Interactive 3D Viz. Tool for NEXRAD II Data
Slide 17/20RCAC – Rosen Center For Advanced Computing
Visualization Images Generated Using Our Tool3D Visualization of 120 NEXRAD II Stations
Scaling – High Resolution Visualization of 3 Radar Stations at different scales ( 1 and 5 )
Rotation of April 10 Tornado3D Visualization of 120 NEXRAD II Stations
Scaling – High Resolution Visualization of 3 Radar Stations at different scales ( 1 and 5 )
Rotation of April 10 Tornado
Slide 18/20RCAC – Rosen Center For Advanced Computing
Hurricane Ike Images (September 14 2008 )
Slide 19/20RCAC – Rosen Center For Advanced Computing
Related Work
• NEXRAD Data Visualization– Integrated Data Viewer (IDV) by Unidata
– National Climate Data Center Java NEXRAD Viewer and Data Exporter
– CRAFT Interactive Radar Analysis System ( Java Viewer )
– LEAD Project ( Gateway )
• Remote Visualization– NanoHub – very small data
– Insley et al. Parallel RayTracing– slow for 3D interactions
• Web Gadgets– Weather.com/Floen.com – simple 2D visualization and
animation of radar data only for one station or whole nation
Slide 20/20RCAC – Rosen Center For Advanced Computing
Summary and Future Work
• Summary of our contribution– A hierarchical and user-oriented design
• rich and easy access to NEXRAD II data for a broad range of users.
– Improved response time and scalability• parallel data pre-processing and partial volume caching
– An integrated end-to-end backend system • radar data retrieval, pre-processing, remote rendering and 3D data
visualization.
• Future Work– Developing optimized data structures by exploiting the spatial
and temporal characteristics of the data
Slide 21/20RCAC – Rosen Center For Advanced Computing
Thank you for your attention!
Q & A