Transcript report on ICWS 07

2007 IEEE Congress on Services
ICWS 2007 and SCC 2007
July 9-13, Salt Lake
City
Program summary






56 Sessions for ICWS 2007 (16 Research Sessions,
31 Application Services and
Industry Sessions, and 9 Work in Progress
Sessions)
32 Sessions for SCC 2007 (10 Research Sessions,
2 Short Paper Sessions, 17 Industry Sessions, and 3
WIP Sessions)
4 IEEE SOA Industry Summit Sessions
2 Keynotes and 2 Plenary Panels, 5 SERVICES
Tutorials
Proceedings: 1234 + 750 + 380 pages !!!
A very selective summary
Tutorials
Tutorial Session 4
Build Your Mashup with Web Services
Ning Yan
Tutorial Session 5
Semantic SOA in Action
P. Deepti B. Majumdar U. Mysore
Build Your Mashup with Web Services
Mashup is presenting new kind of application in web 2.0 world. Mashup is
not simply about the AJAX technologies, rather, it is typically related to
reuse the data and other services from other web side and web
applications. There are many ways to build up the mashup. This half-day
tutorial will focus on using XML and JSON format of data and service and
will introduce the following to the participants:
• What Mashup is about.
• The existing Mashup web sites and applications.
• The Mashup design and architecture which include the client-side mashups
and server-side mashups.
• The existing mashup technologies.
• Case Studies by using JSON services (By using Yahoo’s, Google’s and
IBM’s)
• Case Studies by using XQuery (IBM DB2 V9) by using existing RSS feeds.
At the completion of this tutorial, participants will learn the latest mashup
trends and how to architecture, build and manage the Mashup by using
the existing tools, data and services for their applications.
Semantic SOA in Action:
A Practical Demonstration
This tutorial is intended for researchers and industry practitioners who are
interested in modeling ontologies to support knowledge engineering and
management, database modeling, business process modeling and
enterprise information integration with a focus on practical demonstration.
The tutorial gives a deeper insight for designing of ontologies and
conceptual modeling. It talks about solution strategies for global enterprise
system which provides unified information and agile solution with greater
ease and simplicity. The tutorial also includes demonstration of the
concepts using ontology modeling tools. Global enterprise system needs to
cater the requirements of a constantly changing environment like business
environment changes, user requirement changes and technical
environment changes.
ICWS 2007 Research Track
Session 4, no 1
Semantics in Web Services



Modeling Web Services using Semantic Graph
Transformations to aid Automatic Composition
1 Liu, IBM T.J. Watson Research Center, Hawthorne,
NY
1 Ranganathan, IBM T.J. Watson Research Center,
Hawthorne, NY
1 Riabov, IBM T.J. Watson Research Center,
Hawthorne, NY
p78
Modeling Web Services using Semantic Graph
Transformations to aid Automatic Composition







Main notion is semantic propogation (eg. dependence
of output on input), modeled by graph transformations
only for services that do data processing (stateless –
at least for now)
instead of using ontolgies as in OWL-S
claim it is better for expressing constraints
also compared with SAWSDL
currently no constraints on state – just input/output
used to help connect services as for workflow
Session 4, no 2
Semantics in Web Services



Improved Matchmaking Algorithm for
Semantic Web Services Based on
Bipartite Graph Matching
1 Bellur, Kanwal Rekhi School of
Information Technology, IIT Bombay
1 Kulkarni, Kanwal Rekhi School of
Information Technology, IIT Bombay
p86
Improved Matchmaking Algorithm for Semantic Web
Services Based on Bipartite Graph Matching







Very important/useful background (sect 1+2)
explains the discovery process and limitations (sect 1)
prev work on semantic matchmaking (sect 1) Paolucci [19]
Semantic Matching of Web Services Capabilities + more
outlines OWL OWL-S - use IOPE : Input Output Precond
Effects
mentions DL (Description Logic) used by reasoners
argues with Palolucci's – exact, plugin, subsume
proposes a different algorithm
Session 4, no 3
Semantics in Web Services

Applying Abduction in Semantic Web
Service Composition
1 Lecue, Ecole Nationale Superieure
1 Delteil, France Telecom R&D, France
1 Leger, France Telecom R&D, France

p94

Applying Abduction in Semantic Web Service Composition






background about semantic web services composition
(functional and process)
real life example – phone related services
composition – check if the output of one service matches input
of next (others do this too)
use causal links (from AI planning)
matchmaking – similar basis Paolucci Semantic Matching of
Web Services Capabilities
use concept abduction from [11] Tommaso Di Noia, Eugenio Di
Sciascio, Francesco M. Donini, Marina Mongiello: Abductive
Matchmaking using Description Logics. IJCAI 2003: 337-342
Session 6, no 1
Quality of Service and Security



Probabilistic QoS and soft contracts for transaction
based Web services
1 Rosario, IRISA/INRIA
1 Benveniste, IRISA/INRIA
1 Haar, University of Ottawa, Canada
1 Jard, ENS Cachan Bretagne, France
P126
Probabilistic QoS and soft contracts for transaction based
Web services




mentions orchestration using contract compositions [6, 3]
for the functional part search
but deals with QoS for such compositions. mentions
WSLA (Web Service Level Agreement) – proposed by
IBM [12]
most others deal with hard bounds of eg. response time,
availability, thruput, security (the paper does not deal with
security)
they explain why hard bounds are not good enough – in
practice need "quantiles" eg. 95% of queries answered in
less than 5 msec
Probabilistic QoS and soft contracts for transaction based
Web services (cont.)







here they propose a fully probabilistic approach
(distribution rather than quantiles) for better composition
combine the functional + QoS contracts,
need to be able express concurrency/sequenalization of
orchestration exactly – use math model eg. WFnets
use eg. ORC Misra [23] an academic language for
orchestration Jayadev Misra and William R. Cook
Computation Orchestration A Basis for Wide-Area
Computing also see ORC web site
use monte carlo simulations
survey QoS enabled composition (sect II)
Their tool TOrQuE, examples
Session 7, no 1 Autonomic Web
Service Platform Technology


Automatic Composition of SemanticWeb
Services
1 Kona, University of Texas at Dallas
1 Bansal, University of Texas at Dallas
1 Gupta, University of Texas at Dallas
p 150
Automatic Composition of SemanticWeb Services






need a semantics based approach to make services
accessible by applications rather than humans
mentions effort to built infrastructure for service discovery,
composition etc [17,23,15] –
some are based on semantic web (USDL, OWL-S,
WSML, WSDL-S) others on XML (WSDL) – purely
syntactic
discovery – if a service is not found, try to find 2 or more
that may be composed to synthesize it – a composition
the paper gives (the first?) precise defintion of dicovery
and composition, an algorithm, and a prootype
implementation
based on constraint logic programming
Session 7, no 3 Autonomic Web
Service Platform Technology


A protocol for QoS contract negotiation and
its implementation using Web Services
1 Pouyllau, IRISA/INRIA, France
1 Haar, IRISA/INRIA, France
p 168
Session 8, no 1
Web Services Integration


Rich Services: The Integration Piece of the SOA
Puzzle
1 Arrott, University of California, San Diego
1 Demchak, University of California, San Diego
1 Errnagan, University of California, San Diego
1 Farcas, University of California, San Diego
1 Farcas, University of California, San Diego
1 Kriiger, University of California, San Diego
1 Menarini, University of California, San Diego
p 176
Rich Services: The Integration Piece of the SOA Puzzle





Why Corba did not deliver? tight coupling, heavy
weight of tightly integrating all concerns as all
encompassing middleware
WS – separation of concern.
result – fragmentation of concerns.
challenge: how to integrate the pieces of the puzzle
back into coherent picture, for enterprise scale SOAs.
Introduce Rich Services – basically an architecture
Session 10, no 1 Context-Aware Web
Services Discovery


Handling User Preferences and Added Value in
Discovery of Semantic Web Services
1 Kovacs, MTA SZTAKI, Computer and Automation
Research Institute
1 Micsik, MTA SZTAKI, Computer and Automation
Research Institute
1 Pallinger, MTA SZTAKI, Computer and Automation
Research Institute
p 225
Handling User Preferences and Added Value in Discovery
of Semantic Web Services









A discovery engine
first phase – traditional information retreival (keyword
matching) pre-filtering
second phase - logic based matching, in prolog.
novel prolog style unification of terms
Semantic Web still lacks mechanism for preference
based querying. (Owl-S, based on Description Logic).
match between user goal and web service definition
use Disjunctive Normal Form.
count number of matched terms.
part of a EU project INFRAWEBS
Session 11, no 1 Web Services
Indexing and Discovery


Reputation-Enhanced QoS-based Web Services
Discovery
1 Xu, Queen's University, Kingston, ON, Canada
1 Martin, Queen's University, Kingston, ON, Canada
1 Powley, Queen's University, Kingston, ON, Canada
1 Zulkernine, Queen's University, Kingston, ON,
Canada
p 249
Reputation-Enhanced QoS-based Web Services
Discovery







QoS important when several services provide similar
functionality
UDDI registers cannot provide QoS info.
Authenticity of QoS available info may be questionable
The paper proposes extended UDDI, with QoS, and a
reputation management system.
Some previous work on the use of reputation info
reputation score. less recent data maybe less/more
significant.
Some experiments
Session 11, no 3 Web Services
Indexing and Discovery


Web Service Discovery Using General-Purpose
Search Engines
1 Song, Samsung Information System America
1 Cheng, Samsung Information System America
1 Messer, Samsung Information System America
1 Kalasapur, Samsung Information System America
p 265
Web Service Discovery Using GeneralPurpose Search Engines



used yahoo and google to search for web
services.
yahoo and google were not designed for this,
so try specific approaches (eg. use service
name and/or port type and/or operation
names, with/without additional annotations)
results of experiments . eg. embedding
WSDL into web page enabled both search
engines to find it
Session 13, no 1
Web Services Composition


BPEL4Chor: Extending BPEL for Modeling
Choreographies
1 Decker, University of Potsdam, Germany
1 Kopp, University of Stuttgart, Germany
1 Leymann, University of Stuttgart, Germany
1 Weske, University of Potsdam, Germany
p 296
BPEL4Chor: Extending BPEL for Modeling
Choreographies



BPEL (Business Process Execution Language) is a
language to orchastrate web services into a single
business process.
In choreography, several processes are interconnected &
interaction behavior described from global perspective.
The paper extends BPEL to define choreography.
–
–
–

Participant behavior: control flow dependencies
Participant topology: exisiting participants and their interconnection
using message links
Participant grounding: concrete configuration for data format and
port types
Seamless integration between orchatration and
choreography
Session 14, no 1 Web Services
Modeling and Tooling


Model-driven Composition of Context-aware Web
Services Using ContextUML and Aspects
1 Prezerakos, Technological Education Institute
1 Tselikas, National Technical University of Athens
1 Cortese, Universita' di Roma Tor Vergata
p 320
Model-driven Composition of Context-aware Web
Services Using ContextUML and Aspects




to match user expectations, services should
exhibit some context awareness
but in a of composite service, the context
may have to change.
core service functionality should be separate
from context handling, which cross cut into it.
A proposed solution by Model Driven
approach for design, and AOP for coding.
ICWS Application Services an
Industry Track
Session 1, no 1


A Self-Healing Framework for Web Services
1 Naccache, Arizona State University Tempe
1 Gannod, Miami University, Oxford OH
pp. 398-345
Session 6, no 2






Describing Semantic Web Services: From UML to OWL-S
1 Kim, Yonsei University, Korea
1 Lee, Yonsei University, Korea
pp. 529-536
instead of constructing OWL-S ontology manualy, generate
from UML diagrams.
class diagram represents domain ontology
sequence and activity diagrams represent business process
behavior
XSLT script converts XMI files extracted from UML, into OWL-s
ontology.
Session 11, no 1



Discovering Conversations in Web Services
Using Semantic Correlation Analysis
1 De Pauw, IBM T.J. Watson Research
1 Hoch, IBM T.J. Watson Research
1 Huang, Indiana University
pp. 639-646
An interesting case study related to
orchestration
Session 20, no 2


Raising Programming Abstraction from
Objects to Services
1 Kumar, IBM India Research Laboratory
1 Neogi, IBM India Research Laboratory
1 Pragallapati, IIT Madras, Chennai-600036,
INDIA
1 Ram, IIT Madras, Chennai-600036, INDIA
pp. 864-872
Session 21, no 3





Evolving Existing Systems to Service-Oriented
Architectures: Perspective and Challenges
Hutchinson, Kotonya, Walkerdine, Sawyer
Dobson, Onditi, Lancaster University, Lancaster
pp. 896-903
building SOA system from scratch to replace an
existing system is expensive.
propose how to migrate from an existing system,
resulting in a hybrid system
Session 23, no 1


WS-CDL+: An Extended WS-CDL Execution
Engine for Web Service Collaboration
1 Kang, Southeast University
1 Wang, Southeast University
1 Hung, University of Ontario Institute of
Technology
pp. 928-935
SCC Application and Industry Track
Session 1, no 2


Public Administration Domain Ontology for a
Semantic Web Services EGovernment Framework
1 Goudos, Centre for Research and Technology
Hellas (CERTH), 57001, Thessaloniki, Greece
1 Loutas, Centre for Research and Technology
Hellas (CERTH), 57001, Thessaloniki, Greece
1 Peristeras, National University of Ireland, DERI
Galway, IDA Business Park, Dangan, Galway,
Ireland
1 Tarabanis, Centre for Research and Technology
Hellas (CERTH), 57001, Thessaloniki, Greece
pp. 270-277
Public Administration Domain Ontology for a
Semantic Web Services EGovernment Framework





present a generic Public Administration (PA) domain ontology.
define a formal model for a Public Administration service on the
basis of the Web Service Modeling Ontology (WSMO).
employ the generic public service object model of the
Governance Enterprise Architecture (GEA) providing PA
domain specific semantics.
describe the ontology using the Web Service Modeling
Language (WSML).
This domain ontology is implemented in order to be used in
semantic web services architecture for e-government.