February 04, 2011
- Tim L.
- Jeff DLB
- Matthew Arrott
- John Graybeal
- Navy Considerations
- Data Lifecycle Meeting in CO
- Rich to attend NDBC meeting and start the conversation with Frank. Primary is to establish business interest rather than bring up security concerns.
- Chat with Charly to update the milestones chart.
- Rich will be going to NDBC in two weeks as part of the UAF work -
Original plan to get GTS via LDM - still intended direction, but back-burnered for the time being
[Walt] - Interested in what OOI is doing
Charly: Can provide some material to clarify what he and Matt had done.. Josie + NFRA setup meeting with NDBC
Solicitation of customers + def of customer req. extends further into 2011
Must find resolution on Navy-related talks + community involvement
Agents more stable. Accommodations were made for recent refactorings to system.
Agent communication and technical details consumed more time than planned.
Can we set out to achieve what we need with just the two customers?
Rich: Easiest thing would be to brief to NFRA modeling group. They may have addl streams to add on to the list. Could then get the blessing from the community. Picking up the navy is potentially dangerous - not the most agile community with respect to shifting gears/transitioning
Frank described his problems and it seemed that he would be a great match for our use-case.
Not agile but still have to write custom code from various sources - may appreciate more uniform code base + IT security via OOI.
Rich: Is there someone who can work with us on us. Jeff: IOOS cat will be able to accomplish some, but to get all capabilities should use OOI.
Rich: Would have assumed they had a system in place. [Will make sure to talk to Frank Bub].
Charly: When do we bring in 3rd customer into project (Feb 28) - or too late to incorporate? Chris: Tough to say. Depends on data being requests - whether or not we have support. Different data particularly different sources makes this difficult.
Charly: Can you make sure Jeffs questions answered (via Frank). Chris: IT security is a big deal with Navy, how to we get around this?
Charly: Agreed. If there;s some piece of software thats easy for Wilken/Powell to incorporate, how easy for Frank to do the same. Security is very different:
Rich: Is the pub/sub stuff more intrusive? Is it more difficult to incorp from a security point.
Matt: Its different.. OOI community direction comes from the network rather than the users. Represents less load + helps in that there is one org communicating with them.
Rich: Actual sec risks very different from perceived
Matt: This would be playing to perception more so than reality - at least at this point. Until security mechanisms are in place, treat this as a wash. Believes its more secure, but would be hard pressed to prove this (perception)
Rich: Ex: Replace FTP process to grab model data - discuss communication - sounds dangerous
Matt: For this release, we are already supporting the notion of converting FTP service into a pub/sub service. Going beyond, we will have this capability. This convert should be an external concern to the Navy.
Chris: At this point they would not change anything.
Matt: Franks wider concerns of access to data, is relative to Navy's perception of the data
Jeff: Navy interested in bringing in more data from alt sources. This discussion concerns both Navy getting data and releasing data.
Frank Bub = Navy Oceanographic Office
Navy injesting data
Before second guessing lets setup the call and find out how Frank is getting data from different sites - and what he is permitted to do with data going forward... if he can get data from DAP server, we can setup the dispatcher.. or setup so he interacts with DAP instead of dispatcher being something he deploys (if this is a sec risk)
Rich: With NCEP all data must be in datatank. Jeff: Datatank resp. for running their scripts. There is a lot of ad-hoc coding to acquire data from these source
Matt: Dispatcher rep another script. Question is is the dispatcher ok to run. Chris: Problem is the open connection. Scripts open a temp connection and receives, dispatcher creates an open connection not temp. This is different. There are some similar implementations down there.... so this may be doable
Matt: Through initial conversations these issues will be brought up quickly. Lets find out constraints for the short term
Charly: We were going to send Rich down (on the way to NDBC) to gather this information
Matt: During his time, if there is a smaller conf call, lets address tech issues rather then burdening rich with all information.
Rich: good to go
Matt: Relative to sec questions, very difficult to be all-inclusive in briefing Rich for this.
Dont propose notions that could set off alarms
Rich: Could call Frank - maybe someone better who knows the security const better.
Matt: Doing so we are alerting security concerns without selling the idea first. First establish biz interest and motivation before security concerns are raised. They can inadvertently tie everything down with concerns before the sale is considered.
Matt: Focus on motivating interests. Early discussions were road blocked because of similar issues. - Get biz interests, understand the domain, find what works, heading toward action lands a discussion vetting different options.
Matt: Rich going to NDBC to meet with Frank. If he has more questions than Rich is comfortable responding to we can setup a mini summit similar to that with Brian. This meeting is a good beginning to the conversation
Rich: Knows Frank really well
Charly: He was part of Mass meeting.
Rich: Data life cycle meeting w| Peter Fox. John Graybeal and Michael Meisinger will be attending - March 3rd/4th in Boulder/Denver, CO
Charly: Can't attend. Jeff, if you and Rich check it out, can come to me to discuss. Might be a constraint
Matt: Update to Charly - Bill and he are chatting with Anthony next week.
Charly: Zdenka would like to be part of the conversation before it goes further up.
Matt: Marker is to get Tim Zdenka and Orcutt on the phone
Charly: I'm on standby. Maybe not initial but subs conversation with Tim on [??] level, may be schedule constraints
Charly: Lots of threads to keep our eye on the use cases. Chris, maybe we should chat (soon) update the milestones chart. Ok, if things slide into June, but if there are things we need to add, we need to understand where everything falls
Chris: I'm avail all week.
Data life cycle: Position from Sensor through (filtering, qc, processing, repurposing, dissemination) EVERYTHING