Business Process Management and Semantic Technologies

3/16/2013

1

Business Process
Management and
Semantic Technologies

B. Ramamurthy

3/16/2013

2

Plan for today


Web Interface Design: Lets analyze an
example from Netbeans samples


Mid term review


Chapter 10: relating SOA

BPM

Semantic
Technologies

3/16/2013

3

Introduction


Business Process Management refers to activities
performed by enterprises to optimize and adapt their
business processes.


Business process (BP) is an activity in a company
that uses resources and can involve the activities of
different departments.


BP has been there for sometime but new impetus
has been brought by BPMS, software tools, etc.


Management of BP involves their design, execution
and monitoring.

3/16/2013

4

Business Process
Management


BP Design: capture of existing processes and repositories to
store process models.


BP Execution: uses interfaces and human intervention to execute
the defined processes.


BP Monitoring: involves tracking of individual processes so that
their state can be observed.


Traditional BPM is confined to the boundaries of a business.


The chapter defines a Collaborative Business Model (CBM) that
extends BPM beyond the boundaries of a business.



exploits semantic web services composition engines to do so.


3/16/2013

5

Design of Collaborative
Processes

Modeling

Business

Process I


Modeling

Business

Process II


Lifting

Partner 1

Partner II

SWS

Composition

Alignment

Lowering

Manual

Adaptation

Final

CBP

CBP

Generator

3/16/2013

6

Designing CBP


The authors recommend using (semantic)
Web Services (WS) composition to semi
-
automatically design a Collaborative
Business Process (CBP).


Lifting: transforming BP to WS


Lowering: transforming WS to BP

3/16/2013

7

Semantic Web Services (SWS)
Composition


A semantic WS composition (service
-
enabling) works on semantically enriched
descriptions of Web Services.


3/16/2013

8

Web service definition


Web service description is in a WSDL


WSDL consists of the definition of independent,
atomic and stateless operations


Operations, messages, ports and data types


Data types are in XML schema (XSD)


Messages are input and output messages


WSDL definition will facilitate technical integration of
services.


Observe that WSDL does not have any information
about behavioral semantics of a web service’s
operation

3/16/2013

9

Semantic WS definition


The components of the semantic web service
definition include a formal description of WS
functionality, its inputs and outputs and its
behavioral requirements.


The formal definition of SWS includes an
annotation which is expressed by using
ontology.


Ontology consists of concepts, realtions and
axioms.

3/16/2013

10

SWS definition Languages


OWL
-
S


WSMO (WS Modeling Ontology)


METEOR
-
S


WSDL
-
S

3/16/2013

11

WSMO


Each WS description in WSMO contains a capability


The capability describes WS’s functionality


It is used for discovery and selection of appropriate
services for a specific task as a WSMO goal.


Choreography in WSMO describes behavioral
requirements. This is different from choreography in
WS
-
Choreography standard.

3/16/2013

12

SWS Composition


The business partners participating in the composition provide
SWS : in, out and capabilities in ontological terms.


These are fed into the WS composition engine.


Semantic engine looks for equivalent concepts in the behavior
descriptions and connects them.


After identifying matching concepts, the composition engine
connects fitting input and output.


Result of this process is:


Business processes that contains steps from both partners


Interconnection via mapping activities


Other inputs and outputs that could not be connected as above


Composition is successful when there are no more unconnected
input and output.

3/16/2013

13

Lifting


Transforming process description into format
used by SWS composition engine.


This is achieved by mapping the process
descriptions to the elements of an ontology.


Two parts: lifting in and out messages and
lifting process description



3/16/2013

14

Ontology


A very nice formal description is given in
p.214 and Fig. 10.2


Lets understand this.

3/16/2013

15

Lifting WSDL messages

XSD’s

matcher1

matcherN

Domain

ontology

m
sim

Schema matching

component

m
sim

Aggregation

Alignment

3/16/2013

16

Lifting (contd.)


The architecture of the schema lifting component that creates the
alignment between XML schema and the ontology


This takes XML schema and domain ontology as input and yields
an alignment A
s

o


N matchers are used:


distance matcher


Synonym matcher


Data type matcher


Linguistic matcher


Related entities matcher


The outputs of matching then are aggregated to provide SWS.

3/16/2013

17

Lowering


The CBP is defined as the process steps of
the respective partners plus their appropriate
interconnections.


This step is quite straight forward.


The composition did not alter the original
process structure in the partners.

3/16/2013

18

Execution of CBP


We know how to generate CBP


How to execute CBP?


Fig. 10.5


Fig. 10.6 shows examples of mapping
extraction


Types of rules used in mapping


Move, join, merge, split, replicate

3/16/2013

19

Mapping rules

move

merge

replication

Lifting from source schema

Lifting from target schema

3/16/2013

20

Deployment and execution of
Partner processes


See fig. 10.9


Two businesses execute the CBP designed
by invoking the WS offered by the partners


Each partner has an execution environment
controlling only the execution of their
respective CBP.


Mediation is a important process during
execution.

3/16/2013

21

Overall Procedure


10.9, p.227

3/16/2013

22

Implementation


We studied the design of CBP.


Lets examine the implementation details.


Design using Maestro which is a part of SAP Research business
process management tool suite.


Each partner can create its own business process.


Can be exported, discussed and interconnected with partner’s
processes.


Can be manually adapted (connected if needed)


Workflow orchestrations can be created using a special tool
called ILOG


Lifting


ILOG


Orchestration + Alignment


Lowering


WSDL and WS are deployed using Johnson, Gabriel and
Nehemiah tools


Finalization is needed before actual deployment in runtime.

3/16/2013

23

Demo Scenario


Carrier Shipper Process


Involves three parties: a customer, a shipper
and a carrier


Fig. 10.10
–

10.13


Interfaces, notifications. Contractual
agreements, fulfillments need to be related.

3/16/2013

24

Cost
-
benefit comparison


Agreement on common business terms


Design CBPs using heterogeneous business
processes


Executing CBPs using heterogeneous
message formats


All these save time and effort in manual
processing transformations