1. Part I Web Service Composition
Abdelmounaam Rezgui
Department of Computer Science
Virginia Tech

2. Outline
Summary of the Objectives
Preliminary Approach
Status

3. Objectives Declarative Language for Service Composition
Composers specify what the composite service will do
Automatic Generation of Composition Plans
Formal Verification of Composite Services
Formally measure how close a generated composite service is to the specified one

4. Preliminary Approach Specification Matchmaking
Enables high level descriptions of the desired composition
Task: Develop CSSL (Composite Service Specification Language)
Matchmaking
Generates composition plans
CP = list of component WSs and their interactions

5. Preliminary Approach (cont’d)
Selection
Define QoC parameters (e.g., time, cost, relevance)
Select a plan amongst several composition plans
Generation
Generate a detailed description of a composite service given a selected plan:
List of outsourced services
Mapping between composite service and component service operations
Mapping between messages and parameters
Flow of control and data between component services

6. Declarative Composition of Web Services

7. Specification Phase - CSSL Language -
CSSL: Composite Service Specification Language
<service name=“car broker” category=“brokerage”/> <protocols name=“SOAP”/> <message name=“offer”> <parameter name=“make” type=“string”/> <parameter name=“model” type=“string”/> <parameter name=“year” type=“gYear”/> <parameter name=“mileage” type=“integer”/> <parameter name=“price” type=“float”/> </message>…… <operation name=“receiveSpecialOffers” type=“price sales catalogue” signature=“one way” category=“automobile dealer”> <input name=“offer”/> ……

8. Compatibility Model for Composite Services
Signature
Web Service
One-way
Notification
Solicit-response
Request-response
Operations
Component or Composite
Type
RosettaNet PIP
e.g., price request
Protocols
Category
e.g., SOAP, HTTP
NAICS taxonomy
Parameters
Data Types
XML Schema
Category
Input
e.g., integer
NAICS taxonomy
Output
e.g., insurance
Messages

9. Syntactic Compatibility
Signature Compatibility (operations)
Transport Protocol Compatibility (services)
One-way
Notification
WS 1
WS 2
Solicit-response
Request-response
WS1
WS2
Transport Protocols
Transport Protocols

10. Semantic Compatibility
Operation Semantics Compatibility (operations)
Type (WS1.op1) = Type (WS2.op2)
Category (WS1.op1) = Category (WS2.op2)
Message Compatibility (operations)
WS1.op1.input
data type compatible
WS2.op2.output

11. Status Scanner for WSDL files: Implemented and tested
Syntactic Composability Module
Checks to which extent any two operations of any two Web services are composable

12. Part II Reputation Management
Abdelmounaam Rezgui
Department of Computer Science
Virginia Tech

13. Outline Supporting Privacy Preservation through Trust Approach
Research tasks

14. Trust on the Service Web
Interactions
Involve
Quality expectations
Exchange of sensitive information
Risks
Require
Trust

15. Alternatives for Trust Establishment

16. Proposed Approach Reputation Management Definition: Approach
A service’s reputation within a community of consumers is a perception shared by some or all of the members of that community about that service
Approach
Consumers rate Web services
Consumers collaborate to assess services’ reputation
Trust in services derives from their reputation
Reputation-based service selection and composition

17. Research Tasks Automatic Rating of Web Services
Models and Techniques for Reputation Management
Models for Rating Collection
Reputation Assessment
History-aware Reputation Assessment
Prevention and Detection of Reputation Tampering
Trust-based Selection and Composition

18. An Approach for Automatic Service Rating
Rating Scheme: set of rating rules used to automatically rate services
Generic
E.g., The service exchanges username/password with its consumers only through encrypted communication
Domain-specific
E.g., The pickup location is within 5 miles from the customer’s address
Specification and Interpretation of Rating Rules

19. Models and Metrics for Reputation Management
Centralized, Clique, Arbitrary P2P, Credibility-based P2P
System Metrics
Cohesion degree
Penetrability
Sensitivity
Scarcity
Consumer Metrics
Confidence
Credibility

20. Techniques for Reputation Management
Reputation Assessment in
Regular Environments
Scarce Environments
Discordant Environments
History-aware Reputation Assessment
Prevention and Detection of Reputation Tampering

21. Techniques for Reputation Management
Regular Reputation Environments: dense and relatively concordant
developing distributed algorithms that efficiently collect ratings from consumers and generate accurate reputation values for services
techniques and heuristics to optimize performance and accuracy

22. Techniques for Reputation Management
Scarce Reputation Environments: insufficient density
Statistical techniques
Accurate and fair initialization
assessing the reputation of newly deployed services (i.e., newcomers) for which only a few ratings are available.

23. Techniques for Reputation Management
Discordant Reputation Environments: large numbers of discordant ratings are available for aggregation
Rating Cleansing
Incremental Pruning
Successive iterations produce a relatively concordant environment
Outlier detection techniques
Clustering-merging
Extracts sets of cohesive clusters
Merge clusters to produce concordant environments

24. Techniques for Reputation Management
History-aware Reputation Assessment
Techniques that consider the “effect of history”
Comparison with amnesic techniques
Reputation Prediction

25. Techniques for Reputation Management
Prevention and Detection of Reputation Tampering
Collusion:
occurs when several raters collude with the malicious intention to improve or lower the reputation of one or more services
Whitewashing:
occurs when a service provider suspends and redeploys a service with a different identity with the malicious intention of clearing its reputation history by acquiring a fresh reputation

26. Service Selection & Composition
Trust Derivation
Formally capture the process of trust derivation
Study this process using various distribution for:
Consumer’s confidence in services’ reputation
Consumer’s past experience
Effect of past experience on consumers’ judgment
Reputation-based Selection
Clients are autonomous
May use different criteria for selection
Reputation-based Composition
Static Composition
Dynamic Composition

27. Formal Verification of Composite Service
Measures how a generated composite plan matches a specified composite service
Algebraic approach for Web services
Algebraic Language for the specification of Web services
Calculi for formal verification