Thursday was a half day covering first the cyberinfrastructure and then some discussions of another system that can provide lessons learned to the RSN (reminder: regional scale nodes, the long cabled sensors).
My notes are (of course) at work, but I'll reconstruct some of what I heard. I would recommend interested folks consult the final design document (oh no! it isn't where I found it the other day).
The CI is pretty complicated - in many places it's closer to the cutting edge of science than other parts of the enterprise. The complications include: openness, interactivity, quantities of lots of different types of data, desired latency (or lack thereof), etc. It has to support operation and maintenance of the physical equipment (marine operations), interactive control of some of the devices (particularly gliders and AUVs but other things, too), serving up the data which may include HD video in real time, maintaining provenance in several ways including by keeping level 0 and level 1 data, long term archiving of data, creating a social community around the data to support collaboration among scientists from lots of different disciplines, and finally, supporting education and public engagement.
Their slides will be up in the next week, so you'll see more about the various packages, how they are going to use messaging and cloud computing and more about their network from those.
I'm pasting their topology below, I hope this isn't a problem, if it is, I'll pull it down immediately (e-mail me).
There was a presentation after that by a member of the MARS (Monterey Accelerated Research System) team at the Monterey Bay Aquarium Research Institute (MBARI). MARS is a testbed that supports projects coming out and hooking into a node that is very similar to the ones that will be on the RSN and testing everything from staging to deployment to getting data. They don't have the full cyber infrastructure set up, though. They do have a test tank to practice in before you go out to the site. They have a lot of lessons learned for anyone who is designing sensors for this type of application.
So this was a pretty neat workshop. I'm not sure I got out of it what I was supposed to for work, but I enjoyed it anyway.