Astronomy Networking Needs

1
Grand Challenges in e-Science
Astronomy Networking Needs 5 December 2001 Jim
Kennedy Gemini Observatory Important
Contributions by Dick Crutcher, NCSA, UIUC Tom
Troyland, UKY Arun Venkataraman, NAIC Steve
Grandi, NOAO
2
Research Objectives

• Understand the Universe
• Its Origins
• Current State
• Its Destination
• Origins of Life in the Universe
• Connecting Relativity and QCD, (GUTs/TOEs)
• And Other Modest Goals

3
Research Tools and Venues

• Multi Color Observations, Pictures Spectra
• Radio
• IR
• Visible
• UV
• X-Rays
• Cosmic Rays
• Earth-Based
• Less expensive and maintainable, often remote
• Space-Based
• Removes atmospheric effects, even more remote

4
Data Types and Analysis

• Pictures - 8Kx8K, 16Kx16K, and larger
• Wavelength Spectra - like above
• Temporal Spectra - from time series
• --------------------------------------------------
  
• Raw Data Calibration and Reduction
• Image Enhancement and Reconstruction
• Aperture Synthesis Passive and Radar
• Fourier and Spherical-Harmonic Transforms
• Other Sophisticated Analysis Techniques

5
Typical Issues

• Multi-Site, Multinational Coordination
• Geographically Diverse Communities
• Analysis of Large Data Sets
• Harsh and Remote Environments
• Economical Operations
• Effective Communication with the Public

6
Typical Network Approaches

• Videoconferencing (H.323), Telecollaboration
• Remote Observing, Sea Level or Remote Site
• Automated Observing Sequences
• Data Delivery to Scientists and Archives
• Remote Analysis of Data, Grid Processing
• Network-based Education and Outreach

7
Four Examples

• Gemini Observatory Hawaii, Arizona, Chilean
  Andes
• Seven-nation Partnership
• Two 8m, IR optimized Telescopes, at 14,000 and
  9,000 ft
• Arecibo Observatory Puerto Rico New York
• 1,000 ft Radio Telescope
• NOAO Arizona, New Mexico, Chilean Andes
• US National Facility
• Visible and IR nighttime and solar facilities
• ALMA Chile and elsewhere
• Large multinational Partnership
• 64-Antenna Radio Array at 16,400 ft in Chilean
  Andes

8
Gemini South
Arecibo
Kitt Peak
ALMA
9
Caveats

• Each of these facilities has a great deal in
  common regarding their application needs,
  although the balance between them varies.
• The numbers that follow may be too conservative
  since. In several cases they assume some
  supercomputer-level on-site processing, rather
  than external centers or Grid processing. This
  assumption could prove incorrect.

10
Each Gemini Telescope

• Video/Audio (low latency)
• Real-Time Remote Observing (low latency)
• Real-Time Reduction (quick look)
• Data Delivery to Observers
• Data Archives (CADC)
• Remote Analysis
• Outreach (low latency)
• Bandwidth Requirements
• Today Average 6 Mbps Peak 25 Mbps
• 2005 Average 18 Mbps Peak 50 Mbps

11
Geminis Primary Research Links(Logical
Topography)
12
Arecibo (NAIC)

• Video/Audio (low latency)
• Real-Time Remote Observing (low latency)
• Real-Time Analysis (e.g. pulsars and radar)
• Data Delivery to Observers
• Data Archives
• Remote Analysis (near-real-time and batch )
• Outreach (low latency)
• Bandwidth Requirements
• Today Average 10 Mbps Peak 45 Mbps
  Soon Average 20 Mbps Peak 200 Mbps

13
NOAO KPNO, CTIO, NSO(Each Site)

• Video/Audio (low latency)
• Real-Time Remote Observing (low latency)
• Data Delivery to Observers (several scopes)
• Data Delivery to Internet (10-min turn)
• Data Archives (NVO)
• Remote Analysis
• Outreach (low latency)
• Bandwidth Requirements
• Today Average 10 Mbps Peak 35 Mbps
• Soon Average 25 Mbps Peak 60 Mbps

14
ALMA

• Video/Audio (low latency)
• Real-Time Remote Observing (low latency)
• Real-Time Time Series Reduction (on site?)
• Data Delivery to Observers (all are remote)
• Data Archives (NVO)
• Remote Analysis (supercomputer/Grid)
• Outreach (low latency)
• Bandwidth Requirements
• 2006 Average 32 Mbps Peak 130 Mbps

15
Future Needs

• The rapid growth rate of instrument technology
  and sophisticated data analysis makes future
  bandwidth and QoS requirements difficult to
  predict with accuracy.
• We risk to underestimate them on more than a
  four- or five-year time scale.

16

• End