Web Service Discovery and Selection: Pragmatic Approaches

1
Web Service Discovery and SelectionPragmatic
Approaches

• Natallia Kokash
• PhD student, XX cycle

Tutor Vincenzo DAndrea Adviser Marco Aiello
2
DSSOC people

• Marco Aiello ltaiellom_at_dit.unitn.itgt
• Fabio Casati ltcasati_at_dit.unitn.itgt
• Vincenzo DAndrea ltvincenzo.dandrea_at_unitn.itgt
• Maurizio Marchese ltmaurizio.marchese_at_unitn.itgt
• Ganna Frankova ltfrankova_at_dit.unitn.itgt 3d year
• Natallia Kokash kokash_at_dit.unitn.it 3d year
• GR ltgr_at_dit.unitn.itgt 3d year
• Alexander Ivanyukovich lta.ivanyukovich_at_dit.unitn.i
  tgt 4th year
• Alexander Lazovik ltlazovik_at_dit.unitn.itgt 5th
  year

3
Scope

• Web Service Composition
• Distributed Systems
• Web Service Discovery
• Quality of Service (Security)
• Intellectual Property and Licensing

4
Introduction

• Discovery of web services
• Structural interface matching
• Hybrid methods
• Behavioral interface matching
• QoS Issues
• Web service selection algorithms
• Risk evaluation
• Recommendation systems as a tool for web service
  selection

5
Web Services

• A Web service is
• a software system
• identified by a URI,
• whose public interfaces and bindings are defined
  and described using XML.
• Its definition can be discovered by other
  software systems.
• These systems may then interact with the Web
  service in a manner prescribed by its definition,
  using XML based messages conveyed by internet
  protocols.
• Web Services Architecture, W3C Working Draft 14
  November 2002, from http//www.w3.org/TR/ws-arch/
  on 5th March 2002

6
SOA and Web Services
Service Description
Service Behavior Descriptions
Service Interface Descriptions
Service Quality
Service
Service-oriented application
7
Web Services Discovery

• Matching meeting the functionality required by
  a user with specifications of existing services
• Generic (heuristics, domain-independent
  ontologies)
• Personal (preferences, specific functions and
  patterns for comparing requests and existing
  services)
• Community (domain-specific ontologies)
• Selection choosing a service with the best
  quality among those able to satisfy the users
  goal

8
Thesis objectives

• It is difficult for a user to write a correct
  request.
• Automated semantic matching is not feasible.
• Users are in different conditions.
• QoS characteristics are constantly changing.
• Collective user experience can be used to improve
  service selection.
• Domain specific quality factors should be
  involved in service selection.

9
Service Description

• Web Service Description Language (WSDL)
• Identity unique identity of the interface
• Input/Output the meaning of input and output
  parameters
• Faults specify the abstract message format for
  any error messages that may be output as the
  result of the operation
• Types declare data types used in the interface
  (XML Schema)
• Documentation natural language service
  description and usage guide

10
Semantic Web Services (1)

• Managing End-To-End OpeRations (METEOR-S)
• Semantic Web Services Framework (SWSF)
• Web Service Modelling Ontology (WSMO)
• Ontology Web Language for Services (OWL-S)
• Preconditions a set of semantic statements that
  are required to be true before an operation can
  be successfully invoked
• Effects a set of semantic statements that must
  be true after an operation completes execution
  after being invoked.
• Restrictions a set of assumptions about the
  environment that must be true
• Quality of Service non-functional parameters
  such as response time, execution cost, capacity,
  etc.

11
OWL-S
Resource
Service
provides
presents
supports
Service Profile
describedBy
Service Grounding
Service Model
What does the service do?
How is it accessed?
How does it work?

• Does not provide context identification
• Does not describe objects used by the service but
  not provided by the client
• Does not describe what service does

12
Semantic Web Services (2)

• WSDL-S
• The semantics of services operations are
  directly added to WSDL files
• Easy to deploy and use
• Does not support full features of OWL-S process
  ontology

13
Motivating example
14
WS Matching Algorithm

• Requirements
• Matched advertisements are returned in sorted
  order, according to their degree of match
• For each element a matching confidence is known
  (easy to see where problems may occur)
• Tries to catch semantics

15
Web Service Interface Matching
16
WS Matching Algorithm

• http//dit.unitn.it/kokash/sources

Registry
Parsing
Tagging
Indexing
Query
Ontology
Meta- data

• Tokenization
• sequences of more than one uppercase letters
• sequences of an uppercase letter and following
  lowercase letters
• sequences between two non-word symbols
• Example tnsGetDNSInfoByWebAddressResponse ?
  tns, get, dns, info, by, web, address,
  response.
• Word stemming
• Stopwords removing

17
Structural Matching
Maximum weight bipartite matching Kuhns
Hungarian method (polynomial time) Define overall
similarity score Query type similarity or
inclusion
18
Lexical Matching
Metric
Vector-Space Model (VSM) tf-idf
Semantic

• semantic matching of word pairs
• semantic matching of sentences

VSMWordNet
Seco, N., Veale, T., Hayes, J. An intrinsic
information content metric for semantic
similarity in WordNet, ECAI, 2004, pp. 1089-1090
19
Experimental Results (Test 1)

• 40 web services
• 5 groups

10 40 15 45
VSM
Semantic
20
Experimental Results (Test1)
Average precision
Processing time
21
Experimental Results (Test2)

• 371 web services
• 68 groups

22
Related Work

• Sajjanhar04 Sajjanhar, A., Hou, J., Zhang, Y.
  Algorithm for Web Services Matching, APWeb,
  2004, pp. 665670.
• Bruno 05 Bruno, M., Canfora, G. et al. An
  Approach to support Web Service Classification
  and Annotation, IEEE International Conference on
  e-Technology, e-Commerce and e-Service, 2005.
• Corella06 Corella, M.A., Castells, P.
  Semi-automatic semantic-based web service
  classification, International Conference on
  Knowledge-Based Intelligent Information and
  Engineering Systems, 2006.
• Dong04 Dong, X.L. et al. Similarity Search
  for Web Services, VLDB, 2004.
• Platzer05 Platzer, C. Dustdar, S. A vector
  space search engine for Web services,
  Proceedings of IEEE European Conference on Web
  services (ECOWS), 2005.
• Stroulia05 Stroulia, E., Wang, Y. Structural
  and Semantic Matching for Accessing Web Service
  Similarity, International Journal of Cooperative
  Information Systems, Vol. 14, No. 4, 2005, pp.
  407-437.
• Wu05 Wu, J., Wu, Z. Similarity-based Web
  Service Matchmaking, IEEE International
  Conference on Services Computing, 2005, pp.
  287-294.
• Zhuang05 Zhuang, Z., Mitra, Pr., Jaiswal, A.
  Corpus-based Web Services Matchmaking, AAAI,
  2005.
• Verma05 Verma, K., Sivashanmugam, K., et al.
  Meteors wsdi A scalable p2p infrastructure of
  registries for semantic publication and discovery
  of web services. Journal of Information
  Technology and Management. Special Issue on
  Universal Global Integration, Vol. 6, No.1, 2005,
  pp. 17-39.

23
Hybrid algorithms
Hybridization
Algorithms
Data
Mixed
Switching
Augmentation
Combination
Cascade
24
Hybrid algorithms Experimental results
25
Future work

• Hybrid algorithms
• Rocha, C. et al. A Hybrid Approach for
  Searching in the Semantic Web, International
  World Wide Web Conference, 2004, pp. 374-383)
• Castells, P., Fernandez, M., Vallet, D. An
  Adaptation of the Vector-Space Model for
  Ontology-Based Information Retrieval, IEEE
  Transactions on Knowledge and Data Engineering,
  2007, to appear.
• Empirical evaluation of different algorithms
  using a similar collection of web services

26
Related work

• Syeda-Mahmood, T., Shah, G., et al. Searching
  service repositories by combining semantic and
  ontological matching, International Conference
  on Web Services, 2005, pp. 13-20.
• (1) The domain-independent relationships are
  derived using an English thesaurus (2) The
  domain-specific ontological similarity is derived
  by inferencing the semantic annotations
  associated with web service descriptions.
• better relevancy results can be obtained for
  service matches from a large repository, than
  could be obtained using any one cue alone.
• Klusch, M. Fries, B., Sycara, K. Automated
  Semantic Web Service Discovery with OWLS-MX,
  AAMAS, 2006.
• under certain constraints logic based only
  approaches to OWLS service I/O matching can be
  significantly outperformed by hybrid ones.

27
Composition Patterns

• Sequence
• Loop
• AND split followed by AND join.
• AND split followed by a m-out-of-n join
• XOR split followed by a XOR join
• OR split followed by OR join
• OR split followed by a m-out-of-n join

28
Notation

n

n
s1
29
Behavioral Interface matching

• How to obtain a (composite) service, if there is
  no direct match for a request in current service
  registry?
• Behavioral interface - interfaces that capture
  ordering constraints between interactions.
• BPEL4WS Business Process Execution Language for
  Web services

30
Interface transformation
31
References

• Robert J. Hall and Andrea Zisman, Behavioral
  Models as Service Descriptions, ICSOC, 2004.
• Dumas, M., Spork, M., Wang, K. Algebra and
  Visual Notation for Service Interface
  Adaptation, 4th International Conference on
  Business Process Management (BPM), 2006.
• Benatallah, B., Hacid, M-S., Leger, A., Rey, K.,
  Toumani, F. On automating Web services
  discovery, VLDB Journal, N 14, 2005, pp. 8496.
• Lang, Q.A., Su, St. Y.W "AND/OR Graph and
  Search Algorithm for Discovering Composite Web
  Services", International Journal of Web Services
  Research, 2(4), 46-64, 2005.

32
Web Services Discovery

• Matching meeting the functionality required by
  a user with specifications of existing services
• Generic (heuristics, domain-independent
  ontologies)
• Personal (preferences, specific comparison
  functions)
• Community (domain-specific ontologies)
• Selection choosing a service with the best
  quality among those able to satisfy the users
  goal

33
QoS Issues

• How to define QoS?
• complexity of run-time QoS information
• dependencies among different QoS parameters
• How to specify user preferences?
• How to match user requirements with existing
  services in terms of QoS?
• How to perform ranking of similar services w.r.t.
  to user preferences?
• How to predict QoS factors under certain
  environmental conditions.
• dependencies among different QoS parameters
• relations with contextual factors
• The same questions for composite web services

34
QoS characteristics

• Multidimensionality
• Different QoS driven web service selection
  algorithms
• Subjectivity
• dependence on context, consumer, etc.
• QoS run-time monitoring and analysis on user side
  is required.

35
QoS parameters
36
Linear programming approachZeng et al. 2004
Scaling

• Linear combination of
• price
• duration
• reputation
• success rate
• availability

Weighting
where Wj are user preferences
37
Some questions

• Scaling
• Availability - 100
• 0 0
• 100 1
• Response time
• 0 1
• timeout - 0
• Objective function
• Linear combination - ?
• Can we rely on the preferences defined by a user?
• Which service is better
• Cheap but not reliable,
• Reliable but expensive?
• A service failed but the task should be
  accomplished
• Structure of a redundant composition graph

38
New WS selection algorithm

• Notation
• c composition
• q(si) quality parameter (response time,
  execution cost)
• p(si) probability of success
• qmax resource limit

where

• Time vs. cost
• The basic approach is to take the less important
  parameter as objective function provided that the
  most important criterion meets some requirements.

39
Experimental results
40
Example

• Goal Translate a document from Belarusian to
  Turkish
• Available web services
• Belarusian English (b-e)
• Belarusian German (b-g)
• German Turkish (g-t)
• English Turkish (e-t)
• German English (g-e)
• WS compositions that can satisfy the users goal
• Belarussian English Turkish
• Belarussian German Turkish
• Belarussian German English Turkish

41
Example
g-t
b-g


g-e

e-t

b-e
g-t
b-g


g-e

e-t

b-e
42
Risk management

• Requires assessment of inherently uncertain
  events and circumstances
• Two dimensions
• how likely the uncertainty is to occur
  (probability)
• what the effect would be if it happened (impact)
• Example
• Movie titleRainmaker, formatDVD,
  languagesItalian, English
• Convert DVD to AVI languageEnglish
• SimpleDivX converter time2 hours, language
  Italian
• Impact on time 2 hours are lost

43
Failure risk

• Failure risk considers the probability that
  some fault will occur and the resulting impact of
  this fault on the composite service

where is the probability of the service
failure.
Loss function defines the cost of service
failure (money, time, resources)
44
Scenario
Provider

• Service failures
• Service changes
• Violations of Service Level Agreements (SLAs)
• Absence of alternative solutions (penalties)

Invoke
s0
End-user
45
Failure risk example
46
Related Work

• Zeng 2004 Zeng, L., Benatallah, B., et al.
  QoS-aware Middleware for Web Services
  Composition, IEEE Transactions on Software
  Engineering, Vol. 30, No. 5, 2004, pp. 311327.
• Ardagna 2005 Ardagna, D., Pernici, B. Global
  and Local QoS Constraints Guarantee in Web
  Service Selection, IEEE International Conference
  on Web Services, 2005, pp. 805806.
• Yu 2005 Yu, T., Lin, K.J. Service Selection
  Algorithms for Composing Complex Services with
  Multiple QoS Constraints, International
  Conference on Service-Oriented Computing, 2005,
  pp. 130143.
• Claro 2005 Claro, D., Albers, P., Hao, J-K.
  Selecting Web Services for Optimal Composition,
  Proceedings of the ICWS 2005 Second International
  Workshop on Semantic and Dynamic Web Processes,
  2005, pp. 32-45.
• Canfora 2006 Canfora, G., di Penta, M.,
  Esposito, R., Villani, M.-L. QoS-Aware
  Replanning of Composite Web Services,
  Proceedings of the International Conference on
  Web Services, 2005.
• Martin-Diaz 2005 Martin-Diaz, O., Ruize-Cortes,
  A., Duran, A., Muller, C. An Approach to
  Temporal-Aware Procurement of Web Services,
  International Conference on Service-Oriented
  Computing, 2005, pp. 170184.
• Bonatti 2005 Bonatti, P.A., Paola Festa, P.,
  On Optimal Service Selection, Proceedings of
  the 14th international conference on World Wide
  Web, 2005, pp. 530-538.
• Lin 2005 Lin, M., Xie, J., Guo, H., Wang, H.
  Solving QoS-driven Web Service Dynamic
  Composition as Fuzzy Constraint Satisfaction,
  IEEE International Conference on e-Technology,
  e-Commerce and e-Service, 2005, pp. 9-14.
• Gao 2006 Gao, A., Yang, D., Tang, Sh., Zhang,
  M. QoS-driven Web Service Composition with
  Inter Service Conflicts, APWeb 8th Asia-Pacific
  Web Conference, 2006, pp. 121 132.

47
Quality of Service Issues

• Multidimensionality
• QoS driven WS selection algorithms
• Subjectivity
• dependence on context, consumer, etc.
• QoS run-time monitoring and analysis on user side
  is required.

48
How to define QoS parameters?

• Advertised by providers
• Simple (popular)
• Providers may not advertise QoS information
• Providers are not able to predict QoS in a
  neutral manner
• Providers are interested in overstating the real
  QoS
• Providers do not intend to revise constantly
  advertised QoS
• Not effective and trust-aware.
• Monitored on the client side
• active monitoring and/or explicit user feedback
  (ratings)
• high computational overheads
• Evaluated by a third party
• specialized unbiased agency
• tests web services and publishes QoS data
• expensive and static
• Hybrid

49
QoS Sources
50
Recommendation Systems (RS)

• Examples
• Movies (MovieLens),
• Music (JUKEBOX),
• Books (Amazon),
• Hotels, resorts and vacations (TripAdvisor)
• Types
• Content-based Filtering
• Collaborative Filtering
• Hybrid

51
Content-based Filtering

• Recommendations are based on information on the
  content of items rather than on other users
  opinions
• A buys books on economics
• B is not interested in computer science
• Use machine learning/text mining algorithms to
  create user profiles about user preferences from
  examples based on a description of content

52
Content-based Filtering

• Problems
• Requires content that can be transformed into a
  list of features
• Users tastes must be represented as a function
  of these features
• Unable to exploit quality judgments of other
  users

53
Collaborative Filtering

• Users explicitly assign ratings to items
• Predict rating of a user U for an item I
• Find users similar to U (neighbors)
• Calculate rating of user U to item I as weighted
  sum of ratings given by neighbors to item I

54
Collaborative Filtering

• Problems
• Cannot recommend new items (first-rater problem)
• Random Choose item randomly with equal
  probabilities
• Content analysis Apply previously described
  approach if cannot find similar users
• Filterbots (programs simulating users)
  Constantly do searches and rate items using some
  primitive algorithms
• Cannot match new users they have rated nothing
  (cold start problem)
• Provide average ratings
• User agents collect implicit ratings
• Put users in categories
• Select items for users to rate

55
From recommendations to decision making

• Define a problem
• Return ticket from Trento to Lecce
• Identify alternatives
• By train
• price 100, duration 26 hours, personal
  comfort high
• By plain (Venezia Brindisi)
• price 250, duration 14 hours, personal
  comfort low
• Make the choice
• Train
• Explain the decision
• It is much cheaper

56
Implicit Culture

• Provide actors with suggestions based on
  behavioural patterns extracted from history of
  actions
• Community has knowledge specific to the
  environment community culture
• Encourage a newcomer to behave according to
  community culture
• http//www.dit.unitn.it/implicit

57
Implicit Culture Definitions

• Action something that can be done
• Agent (actor) somebody or something performing
  an action
• Object something that passively participate in
  the action
• Situation a state of the world faced by the
  agent. Includes a set of objects and a set of
  possible actions
• Culture a usual behavior of the group of agents
• Group G group of agents which behaviour is
  observed
• Group G' group of agents who require
  recommendations
• Implicit Culture relation situations in which
  agents of the group G behave similarly to agents
  of the group G'
• System for Implicit Culture Support (SICS) a
  system which tries to establish IC relation

58
System for Implicit Culture Support
Produce a theory about common user behavior
Produce recommendation about action
Stores information about actions
59
SICS Architecture

• The IC-Service is implemented in java and uses
  some libraries
• Can be used in an application in a number of
  ways
• As a java library
• As an EJB component in J2EE environment
• As a web service
• Observations are stored in XML files or in
  database

60
Composer Inductive Module
Exploits the observations and the theory in order
to suggest actions in a given situation.
Analyzes the stored observations and applies data
mining techniques to find a theory about the
community culture
61
IC for Web Service Selection

• How to select a web service with high quality
  suitable for your problem?
• History-based selection
• Quality of Service Quality of Experience

62
Observation of web service invocations

• Actors
• Applications (application name, user name,
  location)
• Users (user name, location)
• Objects
• Operation (operation name, web service name,
  category)
• Inputs/Outputs (parameter name, parameter value)
• Requests (operation names, input/output
  parameters, category)
• Actions
• Bind (timestamp, web service),
• Invoke (timestamp, operation, input),
• Get response (timestamp, operation, output,
  response time),
• Raise exception (timestamp, operation, exception
  type, input),
• Provide feedback (report about contract
  violations, domain-specific QoS parameters),
• Submit query (request, preferences)

63
Example of the theory rules

• Observations
• submit(A, newRequest(categorycurrency))
• invoke(A, http//www.webserviceX.NET/
  CurrencyConvertor conversionRate)
• Theory rules
• submit(_U _Q ) ? invoke (_Y, (category
  extract(category, _Q))) )
• Request
• submit(newClient newRequest(categorycurrency)
  )

64
Future work

• Empirical evaluation
• Customizable similarity evaluation
• Other mining algorithms
• Enrich SICS with semantic matching
• Hierarchy of actions, objects, attributes, etc.

65
Related work

• Blanzieri01 Blanzieri, E., Giorgini, P., Massa,
  P., Recla, S. Implicit culture for multi-agent
  interaction support, Proc. of the Int. Conf. on
  Cooperative Information Systems, 2001, pp. 27-39.
• Maximilien04 Maximilien, E.M., Singh, M.P. A
  framework and ontology for dynamic web services
  selection. IEEE Internet Computing 8(5) (2004)
  84-93
• Manikrao05 Manikrao, U. Sh., Prabhakar, T.V.
  Dynamic Selection of Web Services with
  Recommendation System, International Conference
  on Next Generation Web Services Practices
  (NWeSP'05), 2005,  pp. 117-121.

66
Further information

• Kokash, N. "A Comparison of Web Service
  Interface Similarity Measures", Proceedings of
  STAIRS'06, Riva del Garda, Italy, August 2006,
  pp. 220--231, full paper. Extended version
  Technical Report No DIT-06-025, April 2006,
  University of Trento, Italy.
• Kokash, N., Van den Heuvel, W.-J., D'Andrea, V.
  "Leveraging Web Services Discovery with
  Customizable Hybrid Matching", Proceedings of
  ICSOC, Chicago, December 2006, short paper, to
  appear. Extended version Technical Report No
  DIT-06-042, July 2006, University of Trento,
  Italy.
• Kokash, N. "A Service Selection Model to Improve
  Composition Reliability", International Workshop
  on AI for Service Composition, in conjunction
  with ECAI'06, Riva del Garda, Italy, August 2006,
  pp. 9--14, full paper.
• Birukou, A., Blanzieri, E., D'Andrea, V.,
  Giorgini, P., Kokash, N., Modena, A.
  "IC-Service A Service-Oriented Approach to the
  Development of Recommendation Systems", The ACM
  Symposium on Applied Computing, Special Track on
  Web Technologies (WT), March 2007, to appear.
  Technical Report No DIT-06-044, July 2006,
  University of Trento, Italy.
• http//dit.unitn.it/kokash/