Access Grid: The Good, The Bad and The Ugly

1
Access GridThe Good, The Bad and The Ugly

• Markus.Buchhorn_at_anu.edu.au
• ANU Internet Futures group
• and
• GrangeNet Grid Services Coordinator

2
Overview

• Realtime-Multimedia/IP
• Ye Olde Way
• IP Multicast
• The Access Grid (The Good)
• Who has one, how do you get one
• The Access Grid (The Bad, The Ugly, The Future)

3
Realtime Multimedia/IP
Signal
A
Heart beat
V
Apps
RTP
TCP
UDP
IP
4
Ye Olde Way

• H.320/ISDN, H.323/IP
• Bandwidth Constrained
• Based on ISDN channels, Nx64k
• Site constrained
• One to one interaction
• One A/V stream per site (typically)
• Multi-site interaction requires MCU as
  reflector and transcoder

5
IP Multicast The Good Bits
6
IP Multicast The Bad Bits

• All UDP No TCP
• No reliability
• Of delivery
• Of order
• No QoS
• No tcp-friendly congestion backdown
• No network-layer security
• yet

7
The Access Grid The Good

• www.accessgrid.org - ANL project
• Multicast-based
• Many-2-many interaction
• Multiple A/V streams per site
• Always On approach
• Some Application sharing
• PPT, WWW
• Text Edit, Whiteboard
• No (specified) bandwidth constraints
• Today (h.261 x 2 - 4 audio) 2Mb/s per site

8
The Access Grid Good - 2

• A good/great quality A/V space, if you choose
• Can retrofit H.32x rooms
• Multiple (steerable) cameras
• Multiple microphones, full echo cancellation,
  acoustically-designed spaces
• You need good A/V skills (especially audio)
• A single (more?) large-scale display (3XGA)
• Virtual Venues interaction is done in rooms
• Some recording capabilities
• Eminently Hackable Linux/Windows and open-source

9
An AG Node
Screen
10
Sample AGN Sydney VisLab
11
Sample AGN North Dakota
12
Who, Where, How?

• 140 sites today
• 80 U.S., 30 Europe, 10 Asia
• 3 live in Oz, 15 in plan/build (3 at ANU)
• Mostly at Universities, so far
• Getting one, getting access to one
• GrangeNet participants are supporting their
  roll-out call us!
• Some sites in RD versus/with users

13
AG nodes The Bad, The Ugly, The Future

• A good baseline design
• Wrappers around 5-10 year old tools
• Reasonably standardised distribution (AGiB)
• Virtual Venues
• Some recording capability
• Major problems or shortcomings
• Rudimentary, non-intelligent spaces
• Performance
• Integration
• Applications

14
Advancing AGIntelligent Spaces

• HCI/GUI Issues,
• A/V localisation, Large group representation
• Audio/Video tracking of (moving) targets,
• A/V Synchronisation
• Whiteboards, tablets, laser pointers
• Application Sharing
• One of the hardest things
• Screen scraping vs application awareness
• Centralised and automated node operation
• Debugging (apps, services, networks) and user
  feedback

15
Advancing AGPerformance

• Cluster-driven
• Cameras
• Codecs
• Projection
• High quality A/V
• Resolution (PAL, HDTV, Digital Cinema)
• Latency
• Audio (stereo/DTS/immersive)
• Stereo animation, Stereo video, stereo interaction

• Performance
• Measurement,
• Analysis,
• Tuning
• Stressing
• Reliability/Availability
• Holography
• Haptics
• Non-flat and immersive projection
• Rear-wall projection
• Aids remote interaction

16
Advancing AG Integration

• Intelligent gateways
• H.323, SIP
• Transcoding
• Source selection
• Streaming
• Recording and playback
• Support non-AG spaces, non-GrangeNet bandwidth
• Layered codecs
• Stereo to 2-D

• Security
• Access, encryption,
• Network Services
• Smarter (reliable?) multicast, QoS
• wireless devices
• Additional interaction modes
• Linkage of Venues, Services and applications

17
Advancing AG Applications

• Linkage to other (any!) Grid Services
• computing, data, instruments, visualisation
• Visualisation
• Shared or independent views
• Data browser
• Sharing, discussion, annotation
• Computational Steering
• Instrument Consoles
• Screen Scraping vs application awareness
• Its important!
• Its independent of the space!

18
Advancing AG Other issues

• Physical Security
• Use of public spaces for AGN, tempting equipment
• Use AG equipment (cameras, mikes) for automated,
  intelligent security out of hours.
• Use other sites as your security guard (follow
  the sun security)

19
Summary

• The AG is a (great!) tool, at an early stage
• AG nodes are extensible and flexible
• You can readily bolt in other technologies,
  networks, codecs, room-infrastructure, etc.
• But need to be reasonably standardised,for
  interoperability
• There is a lot of interest in enhancing AG spaces
  and technologies
• Use of AG by user communities (non-CS, non-Net)
  is only just starting