Laws of Relationships Presentation for IRM Summit, Dublin 2014

1. The Laws of
Relationships
(A Work In Progress)
Ian Glazer
Senior Director, Identity
salesforce.com
@iglazer

2. What’s the
problem?

4. firstName
lastName
email
mobile
ou
nickname
title
…
firstName
lastName
email
mobile
ou
nickname
title
…
firstName
lastName
email
mobile
ou
nickname
title
…
firstName
lastName
email
mobile
ou
nickname
title
…

6. Reasonably large
number of identities
with a reasonable
number of attributes

8. deviceID
firmware
deviceID
firmware
deviceID
firmware
deviceID
firmware

10. Unreasonably large
number of identities
with a few
attributes?

12. Reports To
Reports To
Reports To
Works with

13. Reports To
Reports To
Reports To
Owns
Owns
Owns
Works with

14. Reports To
Controls
Reports To
Reports To
Owns
Owns
Owns
Paired
Owns
Gets data
from
Sends data
to
Uses
Works with

15. Reports To
Controls
Reports To
Reports To
Owns
Owns
Owns
Paired
Owns
Gets data
from
Sends data
to
Uses
Drives
Works with
Uses
Constrains
Choice Of
Uses

16. Reports To
Controls
Reports To
Reports To
Owns
Owns
Owns
Paired
Owns
Gets data
from
Sends data
to
Uses
Drives
Works with
Uses
Constrains
Choice Of
Uses
Can send
data to
Riden In
Riden In

17. Unreasonably large number
of relationships between
unreasonably large numbers
of people and things, each
with attributes?

18. Why build laws
in the first
place?

19. • Inform our designs
Test existing solutions
Identify gaps

20. Laws of Identity
(2004)
1. User Control and Consent
2. Minimal Disclosure for a
Constrained Use
3. Justifiable Parties
4. Directed Identity
5. Pluralism of Operators and
Technologies
6. Human Integration
7. Consistent Experience Across
Contexts

21. The Laws Of
Relationships

22. (A Work In
Progress)

23. Acknowledgeabl
e

24. All parties must be able
to acknowledge they are
in a relationship

25. I acknowledge my
relationship with
Twitter Do I
acknowledge
my followers?

26. They will
acknowledge their
relationship with
me
But can I
acknowledge my
relationship with
them?

28. Is this really a Law or a
feature request from the
VRM/PDE/Privacy-types?

29. Actionable

30. Relationships must be
able to carry authorization
data

31. Can perform
actions X, Y, and
Z
Can perform
actions Q, W,
and E

32. Can perform
actions X, Y, and
Z
Can perform
actions Q, W,
and E

33. Can perform
actions X, Y,
and Z
Can perform
actions Q, W,
and E

34. ?
?

35. Constrainable

36. Relationships
must be constrainable

37. With my
permission, it
can report its
location
It can constantly
report energy use
to my power
company
It can only used
by customers
with active
licenses

38. Consent
It can constantly
report energy use
to my power
company
It can only used
by customers
with active
licenses

39. Consent
Consent
It can only used
by customers
with active
licenses

40. Consent
Consent
DRM

41. Contextual

42. Relationships
are contextual

43. Inactive relationships
• None of the parties “use” the
relationship until a condition is
satisfied.
• The set of driver, car, insurer
relationships isn’t “used” until there is
a claim.
• Inert, inactive relationships are
still important because they
provide context
• This widget was made by Yoyodyne.
Drives
Insures
Manufactured by

44. Active Relationships
• Context toggles a relationship
into a usable state
Customer
Owns
Owns
Possesses

45. Context is a requirement
• Related Research:
– Death of authentication and rise of recognition
– Relationship context metadata and the need for durable metadata

46. Immutable

47. Relationships
can be immutable

48. Built by
Built by

49. Provable

50. Relationships
must be provable

51. Mechanism to prove
that a relationship
exists between parties

52. • Single-party asserted
Multi-party asserted
3rd-party asserted

53. Single Party Asserted:
X relates to Y because X says so
I work for
her

54. Multi-Party Asserted:
X relates to Y because X and Y say so
I work for
her
She works
for me

55. 3rd-Party Asserted:
X relates to Y because Z says so
Sally works
for Mary
HR

56. 3rd-Party Asserted:
Does this require other relationships?
HR

57. Revocable

58. Relationships
must be revocable

59. Real-world revocation

60. Real-world revocation
Owns
Owns
Paired
Built By
Acts on behalf of

61. Real-world revocation
Owns
Owns
Paired
Built By
Acts on behalf of

62. Questions that need answers
• Can either party revoke a relationship?
• If I sever a relationship should any party who was part of the
relationship still have access and use of what was shared in the
course of the relationship?
• Does this imply the idea of cascading delete?

63. Scalable

64. • Number of actors
Number of relationships
Number of attributes
Administration

65. Transferable

66. Relationships
can be transferable

67. Temporary Transference
Client

68. Temporary Transference
Client
Delegate

69. Temporary Transference
Client
Delegate
Acts on behalf of
client

70. Temporary Transference
Client
Delegate
Acts on behalf of
client

71. Permanent Transference
Owns

72. Permanent Transference
Owns
Customer Of

73. Permanent Transference
Owns
Customer Of

74. State of transference
• Do we need a system of record for transference state?
• Who would maintain such a system of record?
• Can/should the relationship carry history?

75. Now what?

76. The Laws of Relationships
•Acknowledgeable
•Actionable
•Constrainable
•Contextual
• Immutable
•Provable
•Revocable
•Scalable
•Transferrable

77. Join the Kantara
WG!
https://kantarainitiative.org/groups/irm/

78. IRM WG Goals
• Build a solid set of laws
• Develop an IRM evaluation tool
• Test that tool on well known identity systems
• (Your suggestion here)

79. Go test this!

80. Where should we try and test relationship
management?
• IoT is a natural case
– Industrial settings (factories, planes, etc)
– Citizen (smart homes, sensors in public)
• Familial Relationships
– Insurance
– Healthcare
• Finance
– Complex authorization models
– Regulatory influence

81. Where else can we test this?
• Product architecture
• User stories
• Random strangers on the bus

83. Reports To
Controls
Reports To
Reports To
Owns
Owns
Owns
Paired
Owns
Gets data
from
Sends data
to
Uses
Drives
Works with
Uses
Constrains
Choice Of
Uses
Can send
data to
Riden In
Riden In

84. The Laws Of
Relationship
s