A Formal Model for Classifying Trusted Semantic Web Services

1
A Formal Model for Classifying Trusted Semantic
Web Services

• Stefania Galizia, Alessio Gugliotta, Carlos
  Pedrinaci
• Knowledge Media Institute
• The Open University (UK)
• Asian Semantic Web Conference 2008
• 4th February 2009

2
Outline

• Background on Semantic Web Services and Trust
• Our approach
• WSMO
• Classification
• Trust model
• Conclusions and future work

3
Semantic Web Services

• Use Semantic Web technologies for increasing the
  level of automation within service-oriented
  systems
• Better support key steps such as
• Composition
• Orchestration
• Discovery
• Selection
• Invocation

4
Semantic Web Services

• Use Semantic Web technologies for increasing the
  level of automation within service-oriented
  systems
• Better support key steps such as
• Composition
• Orchestration
• Discovery
• Selection
• Invocation

5
Services and Trust

• In heterogeneous environments, we need to adapt
  service selection based on trust, security,
  performance, etc
• Creating trusted environments requires mechanisms
  to
• monitor, display and analyse information flows
• Provide (ensure) trust in services, according to
  end-users preferences

6
Existing Trust Approaches

• Policy-based
• Binary rating trust, no-trust
• Security certification, trusted certification
  authorities, encryption, authorization
• Reputation-based
• Rating functionally expressed, discrete score
• QoS
• Trusted Third Party-based
• External agent

7
Trust in Context
Virtual Travel Agent

• Am I happy to provide my CC number?
• Are available services secure?
• Can I trust?

8
Trust in Context
Medical Data

• Am I happy with data precision?
• Are available services accurate?
• Can I trust?

9
Framework Overview
Infrastructure for describing, invoking,
discovering/selecting, composing, mediating
SWS Based on WSMO
Stress on context-based and goal-based trust
fits
desires
Goal
Trust Requirements
Trust Guarantees
Ontological Models
WSTO built on a classification library
Classification
WSMO as base formalism for Semantic Web Services
WSMO
10
WSMO Top Level Notions
Objectives that a client wants to achieve by
using Web Services
Provide the formally specified terminology of the
information used by all other components

• Semantic description of Web Services
• Capability (functional)
• Interfaces (usage)

Connectors between components with mediation
facilities for handling heterogeneities
11
Heuristic Classification
12
(No Transcript)
13
Classification library
Classification can be seen as the problem of
finding the solution (class), which best explains
a set of known facts (observables), according to
some criteria.
ltfi, vigt
ltencryption-alg, 3DESgt ltcertification-authority,
verisigngt ltCA-country, USAgt
Observables
ltfi, cigt
ltencryption-alg, symmetric-alggt ltcertification-aut
hority, _gt ltCA-country, americangt
Classification
Candidate Sols.
Solution
WSj
Complete coverage (All features are explained and
none is inconsistent)
Criterion
14
Participant profiles
Participant Profile Trust-Requirements
Trust-Guarantees

• Trust-Guarantees Observables (fi, vi)
• promised features (fp1, vp1), .., (fpm, vpm)
• reliability feature (frL, vrL) execution
  environment monitoring
• reputation feature (fr, vr) external ontological
  statements, for example from other users

• Trust-Requirements Candidate Solutions (fi,
  ci)
• general requirements (fp1, cp1), .., (fpm,
  cpm)
• reliability requirement (frL, crL)
• reputation requirement (fr, cr)

15
WS Profiles
16
Heuristics
17
Classifying Services
ws
Is there any Web Service member of this class?

high-security-encryption
medium-security-encryption
low-security-encryption
low-security-encryption medium-security-CA
low-security-encryption low-security-CA
low-security-encryption high-security-CA
high-security-encryption medium-security-CA
low-security-encryption high-security-CA low-secur
ity-CA-country
low-security-encryption high-security-CA high-secu
rity-CA-country
low-security-encryption high-security-CA medium-se
curity-CA-country
USER-profile-6
18
Trust Model

• Tug(ws) ?(Pg(u), ?g(ws))
• Pg(u) is a function that selects a profile
  associated with the user for the given goal
• ?g(ws) is a complex operator that provides
  information about the ws profile, where ws
  satisfies the goal

19
Trust Model

• ?g(ws) (?pg, ?hg, ?rg)
• ?pg provides the published ws profile
• ?hg assigns values to the ws reliability for each
  feature in the ws profile
• ?rg provides a measure of the ws reputation

20
Summary

• Trust is one of the main challenges for a
  systematic application of Web Services within
  inter-organisational businesses
• Trustworthiness depends on the context
• Our approach enables participants to express
  trust requirements and guarantees
• General requirements/guarantees
• Reliability Monitoring
• Reputation
• Our approach is derived from the heuristic
  classification that has been proved to be widely
  applicable (medical diagnosis, fault detection,
  etc)

21
Future work

• Application of our approach to additional
  non-functional properties (QoS, etc) to obtain
  more adaptive service selection
• New features implementation on IRS-III
• Integration of monitoring for capturing/deriving
  runtime properties
• Reputation Module

22

• Thank you!