A MOTIVATION FOR A MODERN WEB SERVICE
DESCRIPTION
Konstanty Haniewicz
Poznan University of Economics, al. Niepodległości 10, Poznań, Poland
Keywords: Web services, Web service description, Service Oriented Architecture, SOA oriented enterprises.
Abstract: The aim of this work is to present audience with a motivation for research aimed at introducing a new
approach towards a Web service description. The description is referenced mainly at Service Oriented
Architecture oriented enterprises, as they are the most plausible candidates for its employment.
Nevertheless, one might perceive it as a proposition of changes that could have wider reception. The
cornerstone for the research is based on available solutions for a Web service description that are contrasted
with requirements arising from business practitioners. Carried research allowed for stating a set of
requirements that are to be met by a solution that shall improve process of a Web service retrieval. This
works concludes with observations and postulates concerning a modern Web service description.
1 INTRODUCTION
It is believed that Web services are the most
important tool in implementation of Service
Oriented Architecture. At this point every major
production-scale framework for software production
supports Web services. One might risk stating that
Web services became a de-facto standard for
interoperability.
Out of all standards created to enable Web
services, the most important ones are the Web
Service Description Language (WSDL), the
Universal Data Discovery and Integration (UDDI)
and various others devoted to security, quality and
other aspects (involving those that do not abide the
WSDL hegemony). Not all of the crafted standards
became widely adopted by the industry, yet the two
explicitly enumerated have greatly influenced other
initiatives.
Aside from the industry strategies, Web services
became an area of interest for the academia. Very
soon, research works departed from the research of
interoperability to ignite whole new interest in other
features available thanks to employment of Web
services. Researchers observed a great potential of
Web services as an abstraction for certain routines
that could be used to produce highly configurable
software. To enable this to happen, a number of
challenges was identified (Papazoglou, 2007).
First of all, one has to be able to manage his
Web service repository. A considerable number of
works tackled the problem pointing out that the
initial solution (UDDI) is anything but sufficient
(Klein, 2004). Addressing various deficiencies of
UDDI, researchers proposed a number of
enhancements (Hicks, 2007). These enhancements
were realised by building systems that could use
introduced features provided by additions to original
WSDL document. Additions to WSDL documents
could come from semantic extensions (Paolucci,
2002), inclusion of description logic elements not
provided by one of the specific semantic languages
(Colucci, 2003) or processing of WSDL documents
and using the results in the envisioned systems (Al-
Masri, 2009). Over the course of years, semantic
additions became the most prolific area in the
domain of Web services. One has to clearly state
that, popularity of semantic extensions was yielded
by the notion of Semantic net and the idea of
automatic composition of software and data obtained
by enrichment of both, with metadata and
introduction of mechanisms capable of reasoning
over it (Vitvar, 2007). Application of semantics
resulted in rise of Semantic Web Services
(McIlraith, 2001). For a few years SWS become a
term almost universally interchangeable with Web
services.
155
Haniewicz K..
A MOTIVATION FOR A MODERN WEB SERVICE DESCRIPTION.
DOI: 10.5220/0003609801550158
In Proceedings of the International Conference on e-Business (ICE-B-2011), pages 155-158
ISBN: 978-989-8425-70-6
Copyright
c
 2011 SCITEPRESS (Science and Technology Publications, Lda.)

2 CHALLENGES OF A WEB
SERVICE DESCRIPTION
There are a number of approaches to Web service
retrieval. As reported earlier, Web service is in
essence an abstraction for a routine implemented in
a programming language. At this point, one has to
explicitly underline that in most of industry strength
solutions provided by companies such as IBM,
Microsoft, Oracle or SAP a Web service is nothing
more than this abstraction at a level of WSDL
document.
Therefore, at least two main approaches were
crafted. The first one is based mainly on Web
service operation signatures. Main methods of this
approach focus on matching of Web service
operations' signatures. This resembles activities of
earliest research concerned with software
components reuse (Mili, 1997). Signature is not
sufficient to identify purpose and actual activities of
any procedure. Web services were to address this
issue by taking into account preconditions and
effects. Yet, both were broadly disregarded by Web
service community. The sources of disregard are
beyond the scope of this work, nevertheless, one can
reason that its aim were not clear enough to users or
there was not sufficient amount of support at
deployment level.
The second approach takes the observation on
insufficient data stored in routine signature to safely
determine its purpose and uses it as a foundation for
employing any additional data available. This
additional data is stored both in comments inside
routine (and in its neighbourhood) and as actual
names for routine and its parameters.
Both solutions are sure to fail when applied to a
large corpus of Web services (Petrie, 2009).
Therefore, some additional actions must be
undertaken in order to manage data overload.
Organizations that would like to benefit from tools
based on any of further discussed groups of
solutions need to face one or more of following
challenges:
 Semantic information must be attached to Web
services stored in repository
 semantic data must be processed in order to
resolve user queries
 Web services represented only by WSDL
documents thus being signatures are not
sufficient for effective retrieval
 Precise results from queries on Semantic Web
services induce cost of ontology preparation,
weakly prepared ontology results in low
efficiency of query's output
 Extra costs of documenting all items held in
repository, documentation from source code of
abstracted routines might not apply to Web
service itself
 Tagging content with simple categories is too
general and does not save users from further
retrieval by scrutiny
 Cost of search for desired functionality affects
decisions on reimplementation.
Taking into account the above, domain literature
review and results of preliminary interviews with
business practitioners working in IT departments,
there are following postulates that a solution for
Web service retrieval and representation should
have.
Desired solution shall respect current philosophy
employed in design of interfaces search engines.
This is important as users will not be exposed to
another learning curve and will allow for
streamlined adoption among them.
Another important feature that has to be
addressed is description complexity of a single Web
service and additional resources compulsory for
efficient retrieval. Every additional description
element increases complexity. When complexity is
high, provided data can be incorrect due to mistakes
or negligence. Not to mention increased effort of
users generated costs of skill acquisition and cost of
description itself.
Precision is another required feature. Results
must match queries and prevent from presenting user
with abundant information.
Retrieval process shall consume little amount of
time, as business requirements emphasize fast
response. Moreover, interface shall be snappy and
responsive. Solution shall be ready for handling of
hundreds of thousands of Web services.
Performance cannot degrade significantly with
addition of new Web services. Next section shall
cover the available solutions proposed with the
academia. Before that, one has to underline the fact
that a perfect solution shall enable its user to fully
exploit all advantages provided by Web services for
Service Oriented Architecture.
Moreover, anyone that shall try to present a
modern Web service description shall respond to
following questions:
 How one should represent Web service in order
to achieve high precision and manage
description costs at a low level?
 How one should design a system taking into
account all postulates in order to protect it from
performance degradation while increasing
number of stored Web services?
ICE-B 2011 - International Conference on e-Business
156

3 SUMMARY OF A WEB
SERVICE DESCRIPTION
SOLUTIONS
In general, for over 10 years there is no consensus
whether Web services should be enhanced with
semantic annotations or some other method that
increase effectiveness of retrieval should be applied.
Solutions proposed range from simple retrieval base
on processed keywords available in WSDL
documents to those that not only use languages such
as WSML and OWL for description (Maigre, 2010)
of basic functionality but also expression of other
desired by users features.
Most prominent trend in Web service description
is semantic annotation. Greatest benefit of
description with ontology is ability to drastically
increase precision and recall of retrieval process. It's
achieved by unprecedented power to express a
model of some world. Unfortunately, this rebounds
at level of complexity for both world description
builders and model end users.
On the other end of spectrum, there are
description methods that delve into terms used in
WSDL documents (Wu, 2005). In its most naive
form they propose indexing of all available terms
and presenting results as ranked list of WSDL
documents that match query terms. More advanced
solutions allow for pre-processing of terms in order
to filter out possible synonyms and ambiguities. It
was proven that it yields better results than naive
approach. Using traditional retrieval can be
perceived as an iterative process as previous
improvements gain recognition and are included in
later works. Thus when filtering of synonyms, and
partial disambiguation of terms became established
technique, it was enriched by attempts to build
ontology of concepts available in corpus of Web
services accessible to researchers. As a drawback,
one has to highlight that from this point on any
retrieval had to be aided by human operator to some
extent.
Few researchers decided to describe Web
services in alternative manner. Most interesting
method devised is capturing Web service as a
pattern of states and transitions between them
(Rocco, 2005). Flow graphs enabled to perceive
what one can do with given Web service as he is
presented with a list of viable possibilities. One has
to underline that these attempts although tempting,
cannot relieve Web service descriptions builders and
end users from effort spent in learning how to
efficiently model a Web service and later retrieve it.
One last category is hybrid approaches that do
not focus on Web service description technology per
se, but on its scope. This is the most varied category
by far. Common denominator for its members is
inclusion of features that are ignored by previous
categories. As Quality of Service and non-functional
properties are handled in some solutions (WSMO
recognizes non-functional parameters, it realizes
some of QoS with them) distinctive features come
from recognition and addressing issues such as
multiple perspective of Web service, Service Level
Agreement bound to specific operations and Web
services, trust and ability to resolve fuzzily stated
requirements (Cardoso, 2010).
It is crucial, to once more highlight that the
technology in which Web services are described is
of secondary importance. The key, are the features
provided.
4 CONCLUSIONS
One could generalize available solutions in a manner
that varies from those proposed in (Dmello, 2010)
due to the fact that a business user shall not
differentiate between solutions oriented on
functionality or non-functional parameters. They
want a functional entity that shall empower them to
realise their business objectives with minimal effort.
Therefore, there is a number of postulates for a
modern Web service description.
First of all, one would need a purpose statement
of every Web service. This is realized by a few
solutions, one of them is WSMO which allows for
Web service goals. Nevertheless, these service goals
are expressed as capabilities addressing every
element from IOPE quartet. This cannot be universal
solution as a business user is not interested in
preconditions and post conditions. He is interested in
finding a Web service that brings concrete results
and he wants to find it without extra effort in
analyzing ontology interdependencies. It was
observed that tagging systems, based on some
taxonomy are of great interest as they aid to fill
these needs. A Web service is categorized not with
some unrelated terms, but with terms coming from
business user environment.
Thus, Web service description shall take into
account its context. Without it, it is yet another
abstraction layer that can find application possibly
for developers when it's documented. By Web
service context, one can understand its application in
organization. Why was it prepared, and in what
A MOTIVATION FOR A MODERN WEB SERVICE DESCRIPTION
157

terms it was documented. Finally, how it
wasclassified by its builders.
More, Web service can mean different things for
different users. Business user would like to acquire a
building block that can be employed into his
business process and enable him to produce added
value. Developer would like to be able to locate
Web services that can implement some desired
functionality in order to save him from unnecessary
work that could be invested elsewhere. Architect of
organization's system would like to audit state of
affairs and quickly asses whether some functionality
is under or over represented and act accordingly.
External contractor would like to quickly check
whether he can introduce some functionality so that
it can find application in organization and thus bring
him revenue.
Good Web service description shall allow for
multidimensional tagging with a number of
taxonomies. One cannot believe that these
taxonomies should be built automatically. The
process can be aided by traversal of available
documentation and additional input from users.
Ultimately, every taxonomy must be prepared by a
skilled user. Yet when prepared for some compact
area it shall be still easily comprehensible for users
unlike oversized ontologies striving for depicting
domain exhaustively.
One shall believe that only a solution that is able
to harness complexity of semantic technologies and
combines it with good practices taken from WSDL
oriented retrieval can be truly successful.
REFERENCES
Papazoglou M. P, Traverso P., Dustdar S., Leymann F.,
Service-Oriented Computing: State of the Art and
Research Challenges, Computer, vol. 40, no. 11, pp.
38-45, Oct. 2007, doi:10.1109/MC.2007.400
D'Mello D. A., Ananthanarayana V.S., A Review of
Dynamic Web Service Description and Discovery
Techniques, iciic, pp.246-251, 2010 First International
Conference on Integrated Intelligent Computing, 2010
Web Services Description Language (WSDL) v. 2.0,
World Wide Web Consortium (W3C)
recommendation, June 2007; www.
w3.org/TR/wsdl20.
Vitvar T., et al., Semantically-Enabled Service-Oriented
Architecture: Concepts, Technology and Application,
Service Oriented Computing and Applications, vol. 2,
no. 2, 2007, pp. 129–154.
UDDI Version 3.0.1, available at http://uddi.org/pubs/
uddi-v3.0.1-20031014.htm
Petrie C., Practical Web Services. IEEE Internet
Computing 13, 6 (November 2009), 93-96.
McIlraith S. A., Son T. C., Zeng H.. 2001. Semantic Web
Services. IEEE Intelligent Systems 16, 2 (March
2001), 46-53.
Al-Masri E., Mahmoud Q.H., Discovering the best web
service: A neural network-based solution, Systems,
Man and Cybernetics, 2009. SMC 2009. IEEE
International Conference on, IEEE, 2009, p. 4250–
4255.
Crasso M., Rodriguez J.M., Zunino A., and Campo, M.
Revising WSDL Documents: Why and How, IEEE
Internet Computing, vol. 14, Sep. 2010, pp. 48-56.
Hicks J., Govindaraju M., and Meng W., Search
Algorithms for Discovery of Web Services, IEEE
International Conference on Web Services (ICWS
2007), Jul. 2007, pp. 1172-1173.
Klein M. and Bernstein A., Toward high-precision service
retrieval, Internet Computing, IEEE, vol. 8, 2004, p.
30–36.
Mili R., Mili A., and Mittermeir R.T., Storing and
Retrieving Software Components : A Refinement
Based System, vol. 23, 1997, pp. 445-460.
Paolucci M., Kawamura T., Payne T., and Sycara K.,
Importing the semantic web in uddi, Web Services, E-
Business, and the Semantic Web, 2002, p. 815–821.
Colucci S., Di Noia T., Di Sciascio E., Donini F.M., and
Mongiello M., Description Logics Approach to
Semantic Matching of Web Services, Journal of
Computing and Information Technology, vol. 11,
2003, pp. 217-224.
Cardoso J., Barros A., May N., and Kylau U., Towards a
Unified Service Description Language for the Internet
of Services: Requirements and First Developments,
Services Computing (SCC), 2010 IEEE International
Conference on, IEEE, 2010, p. 602–609.
Maigre R., Survey of the Tools for Automating Service
Composition, 2010 IEEE International Conference on
Web Services, Jul. 2010, pp. 628-629.
Wu J., Similarity-based Web Service Matchmaking, 2005
IEEE International Conference on Services
Computing (SCCʼ05) Vol-1, 2005, pp. 287-294.
Rocco D., Caverlee J., and Critchlow T., Domain-Specific
Web Service Discovery with Service Class
Descriptions, IEEE International Conference on Web
Services (ICWSʼ05), pp. 481-488.
ICE-B 2011 - International Conference on e-Business
158