2016-04-19 Meeting Notes

Date

April 19, 2016

Attendees

• Salvatore D'Agostino

• Former user (Deleted)

• Ken Dagg (Unlicensed)
• Former user (Deleted)
• Former user (Deleted)
• Thorsten Niebuhr (Unlicensed)

Goals

• IRM in the Wild Discussion & Examples - Working Through the Principles

Discussion Items

Action Items

Submitted Links

High-level Topics Covered

• IPR & Roll Call
• Meeting Minutes from Last Call
• IRM in the Wild
  
  • Discussed Connected Road Use Case
• Mentioned - Next Step - Start looking at Relationship Manager concept/development

Detailed Meeting Notes

• Kantara IPR - everyone is aware

  Roll call

  Minutes from last meeting are hosted

  • Sal Overview:
  • Went through table on IRM in the Wild - split columns
  • Thoughts are to try couple more use cases
  • Also starting to catch architectural notions
  • May need more discussion on this topic
  • Volunteers to run through use cases?
  • Sal - suggests APIs - use case in the IoT and device management world - sorta covered that already
  • Ken - Sal discussed smart road use case
  • Sal - can do it - interesting use case
  • Ken - sculpt down to a "toll highway"
  • Sal - sure happy to look at connected road and toll use case
  • Ken - feels it would resonate with others if we presented it
  • Ian - interesting and would love to hear it ; Topic at IATP global privacy summit = smart cities - don't have enough to speak to patterns there - clearly government to citizen to platform - convergence is dripping with identity and privacy scenarios - may be a mega header - smart grid is part of that - fertile ground for others
  • Sal - seems like a great one - involved from city surveillance perspective - cue up both of those - play with connected vehicle and smart road to start - then move to smart city;
  • Ian - pointed at privacy perspective - in report out there now - we don't make that much of explicit mention of privacy implications - we are doing a disservice to not call out privacy in our work - tend to want to talk about security benefits; We could benefit from being more explicit - security and privacy are equally important; If when update the report - need explicit mention of privacy benefits;
  • Sal - can capture in architectural notion of this as placeholder; Very good point;
  • Don - fully support of this; Working to make privacy an auditable thing with outside group;
  • Sal - guidelines of surveillance - brings in privacy as well;

  Added "Connected Road" to the IRM in the Wild table:

  • Ian - charge now - who can use your charging unit - whole other can of worms - very much in "UMA" use case as well as IRM
  • Sal - services that are delivered to the vehicle - what's allowed?
  • Ken - topic was connected road - but these are talking about smart car?
  • Sal - indicative of connected road - infrastructure - consider the vehicle platform - then everything else going on (services, etc.) - Infra, Services, Platform = interesting way to look at the use case; Capture layers of the relationships to which they occur; Have to manage all layers;
  • Ken - agree with you - seems to be in the overall picture of the thing - break into smaller chunks and analyze/discuss - wouldn't you break it out into car and road - without worrying about driver of the car, etc.
  • Sal - happy to narrow it
  • Kim - agree - split Connected Road and Connected Car
  • Thorsten - describe a relationship manager
  • Sal - idea that we have IAM and it evolves to IRM - thought is that it is a placeholder for that evolution - puts relationships in context;
  • Ian - thinking about it in terms of needing orchestration and ensuring integrity; Keep finding things that it might be and
  • Identity manager = YES

  Scalable?

  • Sal - initial reaction is YES - but really?
  • Ken - road to car has to be scalable - handling so many cars
  • Ian - are these long term duration or short duration? May have a car again or never again
  • Kim - it is like bandwidth? Burst rates?
  • Sal - short term and not explosive in numbers phenomena - density which is the occupancy x speed - all easily calculated; Traffic management perspective; Transportation network has scalable aspect; Zoom in and zoom out perspective (e.g., few blocks vs. Boston - NYC)
  • Kim - how about partial? Based on road usage
  • Sal - road to car is different than car to road; Works for him - everyone else okay;
  • Ian - asymmetric aspects to this relationship;
  • Sal - architectural notion of asymmetry - directionality;
  • Kim - should we split the column?
  • Sal - smart car to smart road - or anything else?

  Actionable -

  • Sal - has to be

  (Im) Mutable -

  • Ian - context of connected road - may be neither - unless car cares about driver and that relationship is definitely mutable; Road having two ton object on it - definitely immutable; Unsure of data type and content to make statements;
  • Sal - can get into vehicle exclusions - certain areas not driving gasoline tankers;
  • Ken - Ian raises interesting concept - some relationships depend on data being formed as part of relationship - have we explored this?; Data determines mutability;
  • Ian - same page - with that as a frame - what are the examples of the data that is flying around - constrainable to some extent;
  • Sal - toll, balance, picture of vehicle and license plate - examples of data flying around - e.g., license plate lookups to terrorist watchlist; huge amount of data;
  • Kim - is it possible to send out road quality?
  • Sal - example is emergency vehicles turning lights green; Example could be Waze - speed traps, accidents, potholes, etc. are reported; Road conditions - you get feedback on;
  • Ian - push on this one

  Contextual

  • Sal - has to be

  Transferable (Delegation)

  • Sal - not really;
  • Don - makes sense if you have manual control of the car and transfer it to the road;
  • Ian - road to road transfer - that would be a new relationship
  • Sal - goes from road, intersection, etc.; Think it is partial - once automated cars are available;

  Provable

  • Has to be - hard to be anonymous

  Acknowledgeable

  • Fully also
  • Ian - probably is fully - comes back to symmetry question - asymmetrical is amount of data - provable relationship and because it is acknowledgeable - both sides of equation are aware; however each actor is authoritative for is different - equals asymmetry; both parties know they are in a relationship - but eahc masters for different sets of data in the relationship;
  • Sal - for him - acknowledgeable that both parties know that when they get into the relationship; Know there are other drivers, cars, speed limits, etc. which are laws - impact what you can do - consent to letting the road know about me too - agree that it's asymmetrical and has directionality - but none of those are necessarily hidden or not known;
  • Ken - making sense

  Revokable

  • Ian - is there a situation where a vehicle can't transit on a section of road?
  • Sal - yes expired registration, no valid license, etc.; Basic situation - does a car drive on a road - yes
  • Ian - is driver present? Good example
  • Ken - gas guzzling which are not allowed
  • Sal - virtual weight stations are an example;

  Constrainable

  • Partial we already stated
  • Ken - some caused us a great deal of thought - what would have made it less gray
  • Ian - something to help him - know the data in the scenario - what is actually meaningful to each party in the relationship
  • Ken - we have to be very specific - rather than general use cases
  • Ian - doesn't think the answer is yes universally; patterns where data is flowing from one place to another - less relevant than the relationship itself - IoT = data is not needed to prime the conversation; Whereas this use case - the data point is a bit more important; Relationship patterns - relationship is thing which hinges on consideration - other patterns where relationship is secondary to data passed;
  • Ken - feeling uneasy and like we are missing something
  • Sal - interesting thing which pops out - multiple relationship aspects - attribute based control - not combination of ID and what's being managed - multiple of these and all needed to be taken into consideration;
  • Ian - do you mean a whole lot of relationships? Car to road or Road to Car is more meaningful
  • Sal - need to understand "ying and yang" - asymmetry - different data sets - privacy perspective - greatest common denominator - e.g., always accept greatest privacy;
  • Ian - points out that we have been considering single relationships - actor to actor - but we have been looking at just a single edge in the graph - have to start from basics upwards - fast approaching; Good set of use cases; 1) adding actor types - start coming across more realistic use cases, 2) shining a light in the hole which is the IRM Manager; For WG, coming to a point what the big block task is next; Could use more use cases - but writing some of these up would be a good exercise; Updating the principles would be better; Do we want to define Relationship Manager? Or hold off? Question to think about over the next week or two;
  • Sal - we could try and find out - interesting to explore at this point - could be useful for the group to do - if find out we are stuck then do the work to "unstick"