a collaborative research and doctoral educational program based on internet2 inland northwest...
TRANSCRIPT
A Collaborative Research and Doctoral Educational Program Based on Internet2
Inland Northwest Research Alliance (INRA)Dr. Ray Ford - U. Montana
Dr. Gautam Pillay - INRA
Dr. Jack Pelton - Boise St. U.
INRA Collaboration
INRA Role: To support research and educational programs, historically centered around the Idaho National Laboratory (INL/DOE) in Idaho Falls, Idaho
INRA Video
…
[thanks to UM Applications & Media Group]
INRA Collaboration - Relevance to Internet2
The network in general and Internet2 in particular is the “transport base” for all programs
Synchronous and asynchronous delivery of SSGP/PhD level course work and courseware (structured communication)
Growing multi-campus, multi-disciplinary research programs (remote computation, evolving to SCI)
Synchronous communication required to support research and advanced educational programs (unstructured communication)
INRA Collaboration - Remote Video
Five states with five different H.320-dominated distance delivery systems
H.320 system technical compatibility, but fiscal incompatibility => use H.323 (the network is free!)
Site incentive: external funding for facilities and skill development to support move into H.323 “production”
Some more cautious consideration of advanced video technology (e.g. Access Grid Nodes) -- more on that later
INRA Collaboration - H.323 System
9 (then 10) sites: UM, MSU, UA-F, USU, UI, BSU, ISU, INRA, WSU [,WSU-Hanford], most connected through PNWGP
New facilities build out at all sites but “hard” go-live date (Fall 2001) => Started up on time in 8 of 9 temporary facilities, then evolved one by one into permanent quarters
Central H.323 switching from WSU, which thankfully had H.323 expertise (special thanks to R.Cross and J.Hall!)
Central “session capture” at WSU with near real-time conversion to streaming version available on-line -- a very popular feature (repeat special thanks!)
INRA Collaboration - H.323 Results
Tech Coordinator perspective: the result was amazingly good -- decent signal at start-up, with (mostly) consistent improvement as the sites matured
Student/faculty perspective: the result was acceptable at start-up, moving to 8/9 good, but 1/9 unacceptable (following move to a new facility)
Partnership comment: 8/9 acceptable is not good enough
Social comments Who fixes the 1/9 problem? (site, switch, Tech. Coord., a team?) -- a
critical question for program/site operators To students: the streaming version (not real time version) was viewed as
the critical “resource”
INRA Collaboration - H.323 Troubleshooting
We cast our H.323 packets into the air (ether)
They arrive at sites we do know where
We cast our H.323 packets both fat and thin
They arrive at sites who knows when
INRA Collaboration - H.323 Troubleshooting
Observation: H.323 falls between H.320 and TCP/IP Room-to-LAN, LAN-to-edge,
edge-to/from-switch-to/from-edge leaves lots of room for fingerpointing (but who promised QOS anyway?)
H.323 experts are hard to find; TCP/IP experts with time and inclination to think about H.323 problems are even harder to find
Bottom line: IP video troubleshooting can be a great way to “forge” partnerships (in the tradition of very hot fires, lots of pounding, etc.)
INRA Collaboration - Next StepMove to More Advanced Video
Strong desire by all concerned to look beyond H.323 => interest in AGN
UM and UA-F were early AGN adopters, driven by individual researchers. UM, UA-F, U.New Mexico did an “experimental” AGN-delivered course in Fall 2001
Students/faculty: strong interest in (concept of) AGNs Support staff: even stronger prove-that-it-works-first
interest in (realities of) AGN
INRA Collaboration - AGN Evaluation Blues
We (multi)cast our AGN packets onto the net
Most showed up in time (some ain’t here yet)
We (multi)cast our AGN packets, both sight and sound
But re-syncing them up, man its bringin’ me down
INRA Collaboration - AGN Evaluation
Major AGN demo/intro training session in June 2003 INRA site reps mostly collected at UM (Missoula) Two AGNs at UM (i.e., connected by campus network);
single AGNs at each of UA-F, INL, UI, inSORS; bridged H.323 to MSU -- except for “sync” things went well until
Special dose of reality: mid-afternoon lightning strike knocks out some (not-all) campus power, switches, A/C, …
Result: Support for a cautious move to AGN, provided that we address real-time support issues (e.g., how many engineers does it take to reboot an AGN room?)
INRA Collaboration - Current Status
Year #2 (just completing): “smooth” H.323 course delivery
Year #3 (plan): Continue H.323 for course delivery, aggressively use AGN for other communication, develop uses of “shared cyberinfrastructure” to support research
AGN transition: Now rolling out 9 pre-packaged AGNs from inSORS, Inc., plus inSORS central capture/playback; will follow up with training, test sessions, development of “session management” protocols appropriate for available bandwidth, …
INRA Collaboration - What Next
Year #4 and beyond- [ideal] Move to AGN for all remote A/V
- [inevitable] Shift technical focus from A/V support to support for network based resource sharing, i.e., shared cyberinfrastructure, common middleware, federations, etc.