Online Privacy Agent OPA

1
Online Privacy Agent (OPA)

• E-privacy group presentation
• Yingxin He

2
Agenda

• Introduction
• The Goal of Online Privacy Agent (OPA)
• Architecture of OPA
• Handling of Protocols
• Negotiation of Personal Information
• Discussion

3
Introduction

• The Online Privacy Agent (OPA) is the first
  implementation based on P3P and APPEL.
• OPA is implemented at IBM Almaden Research Center
  in 1999.

4
The Goal of Online Privacy Agent (OPA)

• Automatically obtain Web sites policies.
• Automatically evaluate and check these policies
  against users personal preferences.
• Receive assistance during online transactions
  (warnings, recommendations negotiations on sets
  of personal information).

5
Architecture of OPA

• Architecture overview
• Components of OPA

6
Architecture Overview
http//www.almaden.ibm.com/cs/wbi/papers/p3p/Thesi
sJM.pdf
7
Components of OPA
http//www.almaden.ibm.com/cs/wbi/papers/p3p/Thesi
sJM.pdf
8
Handling of Protocols

• Two protocols are used in OPA
• - HTTP and P3P
• P3P was designed to exchange privacy
  information in
• the header of HTTP.
• Early version of P3P supports negotiation, but
  it was
• removed in order to facilitate rapid
  implementation and
• deployment.
• Early version of P3P offers four types of
  messages
• 1. Proposal message 2. Txd message
• 3. OK message 4. Sorry message

9
Negotiation of Personal Information

• The Concepts of Negotiation
• - Terminology
• Information, rules, constraints, and facts
• Rulesets, and their Representation as Tree
• Rule evaluation
• - Counter Offer
• Closest rule, Depth-First-Search
• - Negotiation Strategy
• Summary on Negotiation

10
The Concept of Negotiation (Terminology)

• Information
• P d1, d2, d3,, dn
• a finite set of personal data elements (e.g.
  age, name, etc.)
• Rule
• A pair r (Dr, Cr)
• Dr P, Dr is a set of information
• Cr c1, , cm, Cr a set of constrains
  defined on
• Dr, each constrain in Cr must be met in order
  to get
• access to Dr.

11
The Concept of Negotiation (Terminology)

• Constrain is a relation. It can be unary,
  binary, or n-ary.
• Facts are associated with a request for
  information.
• A pair f (Df, Vf)
• Df ? Ø contains the data elements
• demanded by the requester
• Vf p1,, pl denotes conditions under
• which the elements of Df are requested

12
The Concept of Negotiation (Terminology)

• Rulesets and their Representation as Trees
• - A ruleset Rx r1,,rn
• is a set of rules
• r1 (Dr1, c1, c2)
• r2 (Dr2, c3, c4)
• r3 (Dr3, c3, c5)
• - Tree representation of
• a ruleset

13
The Concept of Negotiation (Terminology)

• Rule Evaluation
• - rule evaluation is a process of matching
• facts against rules.
• - facts must satisfy all the rules constraints
• the requested information in the facts must
• be a subset of the information in the rule.

14
The Concept of Negotiation (Counter Offer)

• Closest rule
• - First, we define a function that measures to
  what
• degree the facts satisfy a rule constraint
• - Second, we define a function that measures
  how
• well the two sets of information overlap

15
The Concept of Negotiation (Counter Offer)

• Closest rule
• - Finally, we define the function ? which maps
  the
• distance between a rule r (Dr, c1, cn)
  and facts
• f (Df, Vf) onto a set of positive integers.
• which means that

16
The Concept of Negotiation (Counter Offer)

• Depth-First-Search

http//www.almaden.ibm.com/cs/wbi/papers/p3p/Thesi
sJM.pdf
17
The Concept of Negotiation (Counter Offer)
Finding the closest rule (DFS, maximum distance
(lt4) for closest rule).
http//www.almaden.ibm.com/cs/wbi/papers/p3p/Thesi
sJM.pdf
18
The Concept of Negotiation (Negotiation Strategy)

• Negotiation strategy represents how a negotiation
  party wants to accomplish its goal in a
  transaction.
• - set maximum distance allowed for the closest
• rule
• - specify weights on constraints.
• - specify a maximum number of negotiation
• rounds allowed during a single transaction.

19
Summary on Negotiation

• A rule specifies constraints on a set of
  information
• A ruleset can be used to verify incoming requests
  for information, and produce counter offer for
  negotiation
• The counter offer is produce from the rule that
  has the minimal distance to the facts.
• Negotiation protocol is not mentioned in this
  paper.

20
Discussion
?