2/17/02 1

Semantic Web Service

Ruoyan ZhangComputer Science Dept, UGA

02/18/2002

2/17/02 2

Talk Outline

Web Service Architecture, Standard, Problem.

Semantic Web Service Specification DAML-S WSMF

Semantic Web Service Discovery Semantic Web Service Composition

2/17/02 3

Web Service

What is Web Service? Web services are the new breed of web application. They

are self-contained, self-describing, modular applications that can be published, located, and invoked across the web. (IBM web service tutorial).

Types of Web Service Simple content-provider implementation (stock quote) Complex process (altering world) (supply chain interaction).

2/17/02 4

Power of Web Services

Universally accessible. It is better than DCOM, RMI,COBRA that

require compatible architectures from all participants to succeed.

Help simplify b2b transactions,Driving down costs and smooth the way to collaborate relationship.

2/17/02 5

Problems Discovery:The programs such as agents have no way

to locate one that will perform a specific function. This process, called service discovery, can happen only when there is a common language to describe a service in a way that lets other agents "understand" both the function offered and how to take advantage of it. Jini…low level, syntactical based, standardized

functionalities. Integration. The development of integrated services

is still largely ad-hoc, time-consuming and requiring a considerable effort of low-level programming.

2/17/02 6

Solutions:

The lack of high-level abstractions and functionalities for web service application has triggered a considerable amount of research. Industry focused on modularization of

services layers-mostly for usability in the short term

Academic Research has been mostly concerned with expressiveness of service description using Semantic Web technology.

2/17/02 7

Web Service Architecture

Figure 1: IBM web services architecture

2/17/02 8

Web Service Standard

Traditional IT integration is build on many systems and protocols.

By using a standardized way to find and use your services(WSDL), you eliminate the need to write new interface each time you integrate with another middleware.

By using standardized Universal Description, Discovery and Integration(UDDI), you do away with manual point to point interface.

By using Web Services Flow Language(WSFL), uses can define the business process and integration with other web services seamlessly.

By using Simple Object Access Protocol(SOAP), you eliminate the need for the old “one adapter per protocol” mechanism.

2/17/02 9

Limits of Industry Effort

Little new really technology, using the standardized taxonomies and vocabularies that exhibit little flexibility and expressiveness and that restricts that usability of web services to human users rather than machine agents. E.g. UDDI: A Web Service user retrieves advertisement

out of registry based on keyword search.UDDI search mechanism relies on pre-defined categorizations through keywords and does not refer to the semantic content of the advertisement. Search only for the on tModelkey,not every attributes. UDDI does not support Semantic description of Services.

WSDL has not precondition, postcondition. It does not support the definition of logical constraints between its input and output parameters,limited in invoking.

2/17/02 10

Semantic Web Service: Go Further Make services computer-interpretable,user-

apparent and agent –enabled. Personalized machine agents to support

automatic Web service discovery, execution, composition and interoperation.

Break out the virtual realm and extend it to into our physical world.

2/17/02 11

Semantic Web Service Specification

DAML-S WSMF-Web Services Modeling

Framework. (D. Fensel,C.Bussler) Process Ontology Other research service ontology on

RDF and DAML

2/17/02 12

DAML-S

A DARPA Agent Markup Language for Services. Semantic markup of Web Services that makes them machine understandable and use –apparent.

BBN Technologies,Carnegie Mellon University,Nokia,Stanford University,SRI International,Yale University

DAML-S 0.6 version released on Dec.2001.

2/17/02 13

DAML-S: Ontology

Service ontology

http://www.daml.org/services/daml-s/2001/10/daml-s.html

2/17/02 14

Service Profile

Specify the intended purpose of the service.

For service discovery. Three types

Provenance and Description

Functionality Description Functional Attributes.

2/17/02 15

Description in Service Profile Service Name Intended Purpose textDescription Provided by Requestedby Actor (requester or provider)

Name Phone Fax Email Physical Address webURL

2/17/02 16

Functionality Description in Profile

Input Output Precondition Effect domainResource

The Class ParameterDescription is used to provide values to inputs and outputs.

parameterName RestrictedTo RefersTo

The Class ConditionDescription is used to provide a conditions and effects.

conditionName Statement*(not done) refersTo

2/17/02 17

Functional Attributes

Address the problem that there are other properties that can be used to describe a service other than a functional process.

E.g. A service may accessed from anywhere on the Internet, it may only applicable to a specific audience. For instance, although it is infeasible to order food from D.C in Atlanta.

GeographicRadius degreeOfQuality Service Parameter communicationThru serviceType serviceCategory QualityGuarentee qualityRating

2/17/02 18

Modeling Service as Process

Two chief components: Process- enables planning,

composition, and interoperation. Process control Model, which allows

agents to monitor the execution of a service request.( not available in DAML-S 0.6).

2/17/02 19

The Process Ontology

Daml-s

DAML-S 0.6

2/17/02 20

The Process Ontology (cont’d)

Three types of Process: Atomic Process

• Directly invocable.• Execute in single step.

Simple Process• Provide a view of atomic

process• Simplified representation

of some composite process(for planning, reasoning).

Composite Process

Control Structure Sequence Split Unordered Split+Join Choice If-Then-Else Iterate Repeat-Until

2/17/02 21

Congo-Purchase: Ontology

B2C-Service

B2C-AccountMgmt

B2C-Lookup B2C-Purchase

Congo-AccountMgmt

Congo-Lookup Congo-Purchase

CP-Profile

CP-Grounding

CP-Model

Process

2/17/02 22

Congo Purchase: ProcessModel

Check-out

Sign-in Create-Account

Load-Prefs Create-Prefs

Select-Payment-Method

Specify-Delivery-Address

Giftwrap

Finalize

One-Step

Locate-Goods Put-in-CartComplexservice

Primitive or complex services

2/17/02 23

CongoPurchaseProcessModel.daml<process>

<name> purchase </name><sequence>

<process> locate_goods (?goods)</process><alternative><alt1><task> One_step </task></alt1><alt2> <sequence>

<task> put_in_cart(?goods)</task><process>

<name> check_out_info</name><alternative>

<alt1> <sequence><task> Sign_in (?user)</task><task> Load_profile(?user.profile) </task>

</sequence></alt1><alt2> <sequence>

<process> Create_Account (?user)</process><process> Create_profile(?user.profile) </process>

</sequence></alt2></alternative>

</process><task> Select_Payment_Method </task>

<process> Select_Delivery_Address </process></sequence></alt2>

</alternative><task> Finalize </task>

</sequence></process>

2/17/02 24

Process Control Ontology

To monitor and Control the execution of a process, an agent needs a model to interpret process instantiations with three characteristics: Provide the mapping rules for various input state

properties to the corresponding output state state properties.

Provide a model of the temporal or state dependencies described by the constructs, sequence…etc.

Provide representation for message about the execution.

2/17/02 25

Resource

Allocation Type resources are allocated to the

activities or process consumabelAllocation ReusableAllocation

Capacity Type DiscreteCapacity ContinuousCapacity

Resource Compostion AtomicResource AggregateResource

2/17/02 26

B2C-Purchase: Grounding• Transport: Secure HTTP

• Protocol: HTTP Forms

• Address: https://buybot.congo.com:4040/initsub.htm

• Type Serializations– ItemDescription (keywords): Set of DAML literals

– PriceRange: pair of monetary units, ISO 5678

– CreditCard: https://transcredit.com/S1.daml#SecureTransferFormat

2/17/02 27

Critique of DAML-S Public Process and internal Process. It

changed from version 0.5 to version 0.6. Still confused on the implementation of simple process.

How to define complex service in Profile? It has no mediation service .

2/17/02 28

WSMF: Goal Strong de-coupling of various

components that realize an E-commerce application.

Strong mediation service enabling anybody in a scalable manner

2/17/02 29

The Web Service Modeling Framework (WSMF)

Ontology: define the terminology that is used by other elements of WSMF specification.

Goal Repository: specify that a client may have in case he consults a web service. Pre-condition Post-condition

Web Services: define various aspects of a web service.

Mediator: bypass interoperability problem.

2/17/02 30

Web Services

1. Name: a unique identifier.2. Goal.3. Precondition and

Postcondition4. Input Data and output data5. Error Data These description elements

still tried a web service as a black box. We need want to have more complex description, failure, concurrent execution, concurrent data input and output, dynamic service binding.

6.Invoked web service proxy.

7.Input ports and output Ports.

8.Data Flow and Control Flow

9.Exception handling10. Error State11.Message exchange

protocol.12.Non functional

properties.

2/17/02 31

Mediator

Mediation of data structures. Mediation of business logics. Mediation of message exchange

protocols.

2/17/02 32

DAML-S vs. WSMF Ontology for Web

Service.Easily to establish the relationship of services.

Public process for monitoring.

Lack of interface definition.

Ontology for concepts.No Service Ontology.

Emphasize the process and mediation in interactions.

Conflicting points in the composition by requestor, since the process is black-box, how requestor know the process?

No semantic support for service composition.

2/17/02 33

Semantic Service Discovery Example: Find a service that sells airline ticket from Beijing to

Shanghai and accepts payment by the MMM Club credit card”

Shoe and Ontobroker. Semantic Search are not enough for the services, especially for the altering-world services since we should consider personal constraints and consequence of services for execution and composition.

2/17/02 34

Service Matching and Discovery

ATLAS: Based on DAML-S by CMU Matchmaking Service by HP Lab, UK Matchmaking using Process Ontology. DReggie: Semantic Service Discovery

for M-commerce application

2/17/02 35

CMU- ATLAS

Based on the DAML-S. Providers advertise the services with

one or more middle agents. Matching Functional attributes.

Determine the applicability of advertisement achieved by performing conjunctive pair-

wise comparisons of the properties.(?) Different type of inference for the different

attributes. e.g. geographicradius and serviceType

2/17/02 36

Matching Functional Attributes

Multiple criteria decision making (MCDM) refers to making decisions in the presence of multiple, usually conflicting, criteria. MCDM problems are common in everyday life. In personal context, a house or a car one buys may be characterized in terms of price, size, style, safety, comfort, etc

Conjunctive constraint method: By setting up a minimum standard for each attribute, the alternative selection or evaluation process is simplified to compare each attribute against its standard. If the standard reflects the decision maker expectations, the obtained solutions are satisfying solutions.

2/17/02 37

CMU- ATLAS (cont’d)

Matching Service Functionalities Service functionalities grouped into

two sets, input and output.

Requester input is Ri, output is RoProvider input is Ai, output is Ao(1) Set Ri is compared to the set Ai, and match is

determined if Ai is subset of Ri.(2) Output sets match when all the elements in Ro

are subsumed by elements in Ao. (i.e)

2/17/02 38

HP-DL Reasoner Description Languages are logic based KR formalisms Based on concepts and roles.

E.g. Man= Human П Male DL System provide efficient inference services

Careful choice of logic/algorithm Highly optimized implementation

DL have been influential in development of Semantic Web

2/17/02 39

HP- DL Reasoner

Based on DAML-OIL, using description logics.

Concept of Match:a service D1 is a match for a service description D2 if there is no contradiction between all of the restrictions in D1 and D2.

Algorithm: The match for service description S Equivalent concepts to S Sub-Concept of S Super-concept of S that subsumed by the Root Sub-concepts of any direct super-concept of S whose

interaction with S is satisfiable.

2/17/02 40

HP-DL Reasoner (cont’d)

Requirements for DL Reasoner for matchmaking SHOQ (D) is the minimum expressive required.(compared to

SHID that does not support individuals or datatypes. Dynamic Ability to deal with multiple interconnected with Tboxes Scalability Persistency Support for DAML+OIL

2/17/02 41

DReggie:

Based on DAML For m-commerce Prolog Reasoner for Semantic Discovery

Representation of a body of knowledge as a set of declarative.

Central to Prolog are concepts from logic:axiom and inference. Prolog makes inference from those axioms about whether certain sentences follow from them

Forward Reasoning and backward Reasoning.

2/17/02 42

Service Composition

Travel Solution Flight Booking Accommodation Booking Tourist attraction searching Bicycle hire and car rental booking Event attendance planner.

2/17/02 43

Problem for Composition How to you know when it is “safe”

to compose A and B?• Output semantics of A match input semantics

of B

The main problem is semantic mapping of arguments

• In most systems to date, this requires human level knowledge.

• If system is purpose-built for composition, semantic mapping can be explicitly coded into service interface.

2/17/02 44

Two Types of Service Composition

Static Service Composition: Plan of the composite service is known a priori. E-flow is platform for the specification, enactment

and management of composite e-services.

Service Discovery Pattern: What type of information is needed to identify

service components at run-time? How to integrate the component services in the

provision of this composite service?

2/17/02 45

Service Discovery Pattern(cont’d)

Solution: An automated service discovery:given that component

services are described using software interpretable information(ontology), this facility provides means(service discovery engine) to locate component services based on constraints over their meta-data.

A composite service specification allows component services to be automatically discovered and integrated.

Implementation. CMI(placeholder->selection policy) Push-Community.(WebBIS)

2/17/02 46

Push-Community:

Describes the capabilities of a desired service without referring to any specific service.

A services can register with it to (fully or partially)offer the desired operations. This involves the definition of the mappings between operations defined in the community and those defined in the actual services.

Push Community defines a request for services. The task of composing a complex service is gracefully

distributed by enabling each provider to contribute to one or more communities without knowledge of the details of other participating services.

Abstract Service Ontology, instead of the concrete service in the query.

2/17/02 47

Composite Service Execution (I)

Service Execution with Central Authority Service provider host composite scheduler.

• Initiating the execution of the components of services.

• Receives and processes service request• Handling and processing data according to the

data semantics of the composite service. Implementation: Eflow has no recursive

functions.

2/17/02 48

Composite Service Execution (II)

Peer-To-Peer Execution Pattern. limit as much as possible the number of messages exchanged

during the execution of a composite service to avoid this potential bottleneck

Execution not only depend on central scheduler, but slow on the coordinator hosted by each of the providers.

Implementation Aspects: What are the preconditions for triggering the execution

of a service? When the service execution is completed, what are

service that may potentially need to be executed next? Known implementation:

Selfserv

2/17/02 49

Semantics?

Specify the control and data flow semantics of the new services based on elementary services or other composite services.

Dynamic Service discovery Abstract Service Specification.

2/17/02 50

DAML-S:Agent Technology for Service Composition

Objective:develop agent technology that will perform these types of tasks automatically by exploiting markup(knowledge representation) of Web services and markup of user constraints.

Planning is computationally intensive. reusable generic procedures with customizing

user constraints Theme: Usability and customization.

http://semanticweb2001.aifb.uni-karlsruhe.de/slides/mcilraith-semweb01.pdf

2/17/02 51

Generic Procedure

Agents are tasked using high-level, reusable”generic procedure”

“The what not necessarily the how” E.g. Book-transportation(origin,des-d,date-r,purpose) pick-one-of Book-plane(origin,dest)| drive(origin,dest)|Book-train(origin,dest) end pick end

Generic Procedures stored in mark-up ontologies using the same DAML-S like markup.

http://semanticweb2001.aifb.uni-karlsruhe.de/slides/mcilraith-semweb01.pdf

2/17/02 52

Realizing Generic Procedure in Golog

Golog: a high-level logic programming

language, for specification and execution of complex actions in dynamical domains.

Built on the top of situation calculus, a first-order logical language for reasoning about action and change.

Procedure constructs: Sequencing If-then-else Nondeterministric choice While-do, etc.

SRent-car

Get-plane-sche

Buy-plane-ticket

http://semanticweb2001.aifb.uni-karlsruhe.de/slides/mcilraith-semweb01.pdf

2/17/02 53

Customizing User Constraints Generic Procedure can be further constrained by DAML-S defined user constraints

Personal constraints/preferences Group constraints, or Instance-specific constaints

E.g. Bob would like to drive if the driving is less than 3 hours KSL business air travel should be on American carrier.

http://semanticweb2001.aifb.uni-karlsruhe.de/slides/mcilraith-semweb01.pdf

2/17/02 54

Deductive Instantiation Agent’s KB is automatically constructed relative to the

generic procedure and user constrains. Deductive machinery instantiates the generic procedure

wrt, constraints and world states to generate Web service requests that the agent broker execute.

Middle ground interpreter balances information gathering services with delayed execution of world-altering services

Agent’s KB is updated by Web Service response.

http://semanticweb2001.aifb.uni-karlsruhe.de/slides/mcilraith-semweb01.pdf

2/17/02 55

Agent Technology Framework

http: / /semanticweb2001.aifb.uni-karlsruhe.de/slides/mcilraith-semweb01.pdf

Web procedure Ontology

Web Service Ontology

Agent Broker

Personal and company Constraint

DAML-enabled web pages

E-mailKB

2/17/02 56

Comments on DAML-S

What is difference between the services and generic procedure?

It is important to discover the service at run-time .

Planner?

2/17/02 57

Reference 1.Web Service: Semantic Web Service Architecture – Evolving Web Service Architecture-

Evolving Web Service Standards toward the Semantic Web. T Sollazzo, S Handschuh, S.Staab, M.Frank.

2.Semantic Web:        The Semantic Web: T by TIM BERNERS-LEE, JAMES HENDLER and

ORA LASSILA         Agents and the Semantic Web. James Hendler.        Semantic Web Service. S.A McIlraith, T.C. Son, HongleiZeng. 3.Web Service Specification and Modeling Framework: DAML_S: Semantic Mark Up for Web Services: The DAML Services

Coalition: DAML-S The Web Service Modeling Framework WSMF, D. Fensel and C. Bussler

WSMF

2/17/02 58

Reference (cont’d)4..Web Service Discovery and Matchmaking:  A Semantic Web Approach to Service Description for Matchmaking of Services: Description Logics for Matchmaking of Services Searching for Services on Semantic Web using Process Ontologies, Mark Klein,

Abramham Berstein. Dynamic Service Matchmaking Among Agents in Open Information Environments.

Katia Sycara, M.Klush, S.Widoff. Advertising and Matching DAML-S Service Description. T. R. Payne, M. Paolucci,

K.Sycara Dgeggie: Semantic Service Discovery for M-Commerce Applications. D. Chakraborty,

F. Perich, S. Avacha, A.Joshi Web Service and Dynamic Discovery. S. Vinoski  3.Semantic Web Service Composition:Adapting Golog for Programming the Semantic Web. S. McIlraith, T.C. Son.Adaptive and Dynamic Service Composition in e-Flow. HP Laboratory.Towards Patterns of web Services Composition. B Benatallah, M.Dumas, M-c.Fauvet,

F.A.RabhiWeb Services Flow Language. IBM Software Group.