Agent%20Paradigm:%20a%20Promising%20Approach%20to%20Enterprise%20Application%20Integration

1
Agent Paradigm a Promising Approach to
Enterprise Application Integration

• Paola Turci
• turci_at_ce.unipr.it

2
Outline of the Lecture
? The Context ? Consolidated Approaches Open
Issues ? Research Directions ? An Agent-Based
SOA
?
3
Todays Enterprise Application Issues

• Ability to change IT quickly to fit business
  needs
• To improve responsiveness and efficiency
• Demand for high-levels of interoperability
• Organizations feel the need to support
  interoperability between separately designed
  systems
• Most enterprises are very heterogeneous
• Organizations want applications to have broader
  reach

4
Business Performance Management Institutes
Survey - 2006

• 11 percent of executives say they're able to keep
  up with business demand to change
  technology-enabled processes 40 percent of
  which, according to the survey, are currently in
  need of IT attention
• 36 percent report that their company's IT
  departments are having either "significant
  difficulties" (27 percent) or "can't keep up at
  all" (9 percent)

5
What Characterizes Enterprise Applications?

• Usually
• Involve persistent data
• There is a lot of data
• Many people access data concurrently
• A lot of user interface screens
• Need to integrate with other enterprise
  applications

6
Enterprise Application Integration

• The task of making disparate applications work
  together
• EAI challenges
• Applications can run on multiple computers, which
  may represent multiple platforms, and may be
  geographically dispersed
• Networks are slow and unreliable
• Applications may run outside of the enterprise by
  business partners or customers
• Applications are quite dissimilar
• Changes are inevitable

7
Enterprise Application Integration (II)

• Criteria that should be considered when choosing
  and designing an integration approach
• Application coupling
• Intrusiveness
• Technology selection
• Data format
• Data timeliness
• Data or functionality
• Remote Communication
• Reliability

8
Outline of the Lecture
? The Context ? Consolidated Approaches Open
Issues ? Research Directions ? An Agent-Based
SOA
?
9
Integration Approaches

• Four main integration styles consolidated
• File Transfer (70s) applications share files
• Good physical decoupling
• Language and system independent
• No extra tools or integration packages are needed
• An important decision is what format to use (XML)
• Drawbacks
• Effort is required to produce and process files
• Need to agree on the filename, location and
  format, the timing of when it will be written and
  read, and who will delete the file
• Possible semantic dissonance
• Systems can get out of synchronization

10
Integration Approaches (II)

• Shared Database (80s) applications share the
  same database schema, located in a single
  physical database
• Use of SQL-based relational databases
• Consistent data
• Drawbacks
• Integration of data
• Difficult to find a common representation
• Database may become a performance bottleneck

11
Integration Approaches (III)

• Remote Procedure Invocation (90s) Each
  application seen as a large-scale object or
  component with encapsulated data
• The communication is synchronous
• Goal
• Make RPC look as much like local PC as possible
• Advantages
• Data exchanged only as needed
• Integration of business functions, not just data
• Weaknesses
• Big differences in performance and reliability
  between remote and local procedure calls
• Latency
• Lack of control over other systems
• Applications tightly coupled as a local method
  call

12
Integration Approaches (IV)

• Messaging applications publish messages by
  using a common messaging system
• Agreement on the message format.
• The communication is asynchronous
• Pros
• Reliability
• Loose coupling
• But fewer assumptions mean that more is left to
  do for the developer
• Correlation of related messages (e.g. request and
  response)
• Maintaining state
• Determining order of events, to re-establish the
  message sequence
• Complex programming model figuring out what to
  do next
• Main drawback
• Individual systems are wired together via a
  message flow

13
remarks about coupling

• Coupling is not inherently good or bad
• Coupling is a measure of dependency between
  applications
• Technology Dependency
• Location Dependency
• Temporal Dependency
• Data Format Dependency
• Tightly Coupled Systems
• Make many assumptions about each other
• Well suited for internal communication inside of
  an application
• Well suited for near communication, with
  control over both sides of the interaction
• Generally more efficient, easier to develop and
  debug

14
Forrester Research

• SOA's ability to save IT costs and, more
  importantly, build business flexibility has
  driven broad adoption 53 of enterprises are
  using SOA now or will use it by the end of 2006,
  and nearly one-half of large enterprises are
  using SOA for strategic business transformation.
  SOA is applicable to a broad range of business
  and technical scenarios, and every organization
  should investigate SOA and learn where and how it
  can benefit them
• May, 2006Topic Overview Service-Oriented
  Architectureby Randy Heffner, Larry Fulton
• An independent technology and market
  research company that provides its clients with
  advice about technology's impact on business and
  consumers

15
CIO Magazine

• According to a recent Forrester Research
  survey, 46 percent of large-enterprise SOA users
  (and about 27 percent of SOA users at midsize and
  smaller enterprises) said they're using SOA to
  "achieve strategic business transformation."
  Surveys from other research companies report the
  same enthusiasm
• BUT
• SOA is far from being a proven concept (only
  16 percent of companies in the Aberdeen survey
  have more than 24 months of experience with SOA
  technologies), and the companies that have had
  the most success with it so far are those that
  always have success with technology big
  companies with big IT budgets whose business is
  technology-based (think telecom and financial
  services). They also tend to have supportive,
  technologically sophisticated business leaders
• The Truth About SOA
• June 15, 2006

16
What is SOA?

• A form of distributed systems architecture,
  typically characterized by the following
  properties  
• Logical view The service is an abstracted,
  logical view of actual programs, databases,
  business processes, etc., defined in terms of
  what it does
• Message orientation The service is formally
  defined in terms of the messages exchanged
  between provider agents and requester agents, and
  not the properties of the agents themselves. A
  key benefit of this concerns so-called legacy
  systems
• Description orientation A service is described
  by machine-processable meta data. only those
  details that are exposed to the public and
  important for the use of the service should be
  included in the description. The semantics of a
  service should be documented, either directly or
  indirectly, by its description.  
• Granularity Services tend to use a small number
  of operations with relatively large and complex
  messages.
• Platform neutral Messages are sent in a
  platform-neutral, standardized format delivered
  through the interfaces (XML)

17
what is the point?

• Most of the technology and market research
  companies, which provides their clients with
  advice about technology's impact on business and
  consumers, agree on the fact that the adoption of
  a SOA paradigm is strategic and should be part of
  the most forward-looking software projects.
• Nevertheless
• the paradigm shift is still quite challenging!!!

18
Widgets Gadgets 'R Us An Example

• An online retailer that buys widgets and gadgets
  from manufacturers and resells them to customers
• Requirements
• Take Orders customers can place orders via Web,
  phone, or fax
• Process Orders processing an order involves
  multiple steps verifying inventory, shipping the
  goods, and invoicing the customer
• Check Status customers can check the order
  status
• New Catalog suppliers update their catalog
  periodically. WGRUS needs to update pricing and
  availability based on the new catalogs
• Announcements customers can subscribe to
  selective announcements from WGRUS
• Testing and Monitoring operations staff needs to
  be able to monitor all individual components and
  the message flow between them

Source G. Hohpe, B. Woolf Enterprise
Integration Patterns Addison Wesley
19
Widgets Gadgets 'R Us (II)

• WGRUS IT Infrastructure

Source G. Hohpe, B. Woolf Enterprise
Integration Patterns Addison Wesley
20
Taking Orders

• A message-oriented middleware solution to
  streamline the order entry process

Source G. Hohpe, B. Woolf Enterprise
Integration Patterns Addison Wesley
21
Processing Orders
Source G. Hohpe, B. Woolf Enterprise
Integration Patterns Addison Wesley

• A typical implementation of a distributed
  business process
• Main disadvantage of such a solution
• Individual systems are wired together via a
  message flow

22
Process Manager

• Individual systems turned into shared business
  functions can be accessed as services
• Increases reuse and simplifying maintenance
• Interlinked by statically or dynamically defined
  workflows
• Services orchestrated via a Process Manager

Source G. Hohpe, B. Woolf Enterprise
Integration Patterns Addison Wesley
23
Process Manager (II)

• To turn the WGRUS IT infrastructure into an SOA
  it is necessary to add facilities to look up
  ("discover") a service from a central registry
• In order to participate in this SOA, each service
  would have to provide additional functions
• e.g. to expose an interface contract that
  describes the functions provided by the service
• Each request-reply service also needs to support
  the concept of a return address

24
the Vision
Legacy SW
Web Application
HTTP
CORBA
Business Process
ODBC
Java Application
RMI
Data Base
WS-BPEL
.NET REMOTING
JCA
ERP, CRM
.NET Application
Web Service
25
XML SOAP WSDL UDDI

• Rely on static descriptions of service
  interfaces, forcing consumers
• to find and bind services at design time
• Do not address runtime service selection based on
  a dynamic assessment of nonfunctional attributes
• They guarantee syntactic interoperability, but
  they fail to provide semantic operability

26
Open Issues

• This innovative idea brings with it new
  outstanding opportunities but also new great
  issues
• How to efficiently discover Web services
• How to allow and facilitate their composition

27
Outline of the Lecture
? The Context ? Consolidated Approaches Open
Issues ? Research Directions ? An Agent-Based
SOA
?
28
Agent Community

• Evidence from several research studies that
  agents could represent a suitable technology
  which can be used to meet the performance needs
  for innovative business applications
• Several papers attest the current interest in
  using agents for developing e-business
  applications, business process management and
  enterprise application integration
• Different works have shown how agent technology
  can be leveraged if used together with other
  technologies semantic Web, Web services, rule
  engine and workflows

29
(No Transcript)
30
OMG SOA

• 2006, OMG launched a SOA SIG.
• Primary goals
• To support an MDA approach to SOA that links
  architectural,  business and technology views of
  services, including Business Process  Management
  (BPM) and Event-Driven Architecture (EDA).
• Identify and foster development of OMG modeling
  standards for SOA that integrate with and
  complement standards developed by other
  organizations such as W3C, Open Group and OASIS.
• To improve awareness and understanding of SOA by
  OMG members.
• To coordinate SOA related efforts within OMG.
• But little has been done!!!

31
FIPA - Agents and Web Services Interoperability
Working Group

• Since 2005, FIPA is an IEEE Computer Society
  standards organization
• Interoperability and standardization efforts
• The primary objective of AWSI WG is to fill the
  interaction gap between agents and web services.
  Agents should be able to locate and interact with
  web services seamlessly and vice versa
• But most of the work has to be done!!!

32
Research Directions for Service-Oriented
Multiagent Systems

• Four major trends will drive SOC and MAS research
  in the next decade
• Online ontologies shared representations are
  emerging as models for real-world entities
• Ubiquitous computing
• Computational behaviours provided in the form of
  Web services and service architecture
• Widespread availability of sensors and effectors
  control of physical world
• SOC represents an emerging class of approaches
  with MAS-like characteristics for developing
  systems in large-scale open environments

Huhns, M., Singh, M., Burstein, M., Decker, K.,
Durfee, E., Finin, T., Gasser, L., Goradia, H.,
Jennings, N. R., Lakartaju, K., Nakashima, H.,
Parunak, V., Rosenschein, J., Ruvinsky, A.,
Sukthankar, G., Swarup, S., Sycara, K., Tambe,
M., Wagner, T. and Zavala, L. (2005) Research
directions for service-oriented multiagent
systems. IEEE Internet Computing 9(6) pp. 52-58.
33
W3C Workshop on Frameworks for Semantics in Web
Services June 2005

• Two concrete proposals
• Elicit some concrete challenges (pain points)
• Bring together the developers of the more
  comprehensive technology efforts to define simple
  specifications for the low-hanging fruit that
  is common to all of their approaches
• Semantic annotation in WS descriptions (such as
  typing of inputs and outputs, expression of
  preconditions and effects of primitive actions)
• A simple incremental set of extensions to WSDL to
  indicate references to these elements
• two years after SAWSDL
• Hierarchical classification of services for
  purposes of advertising and discovery
• Expression of non-functional properties such as
  QoS
• Possibly simple uses of rules in expressing
  policies of contractual commitments
• two years after Web Services Policy

34
Outline of the Lecture
? The Context ? Consolidated Approaches Open
Issues ? Research Directions ? An Agent-Based
SOA
?
35
the new Vision
Legacy SW
Web Application
HTTP
CORBA
ODBC
Java Application
RMI
Data Base
MAS
.NET REMOTING
JCA
ERP, CRM
.NET Application
36
Agents Role

• The central role that agents should play in a SOA
  scenario is
• To efficiently support distributed computing
• To allow the dynamically discovering and
  composition of Web services
• To be successful, it is crucial
• to appropriately engineer and integrate agent
  technology with other
• strategic technologies

37
Ontologies

• An explicit specification of a
  conceptualisation Gruber93
• Research community contributions have been mainly
  devoted to cope with four different issues
• Formal definition of a standard language for
  expressing semantics on the web (OWL)
• OWL Group to refine and extend Web Ontology
  Language
• Development of tools to engineer ontologies
• Two co-existing realities
• Semantic web
• Object-oriented systems
• An ontology representation more in line with the
  OO data model

38
Ontologies (II)

• Development of ontological supports for MASs
• A communication support to perform the proper
  semantic checks on a given content expression
• e.g. the JADE ontological support
• Definition of specifications for semantic Web
  services
• Goal providing a semantic layer based on the WS
  standards
• relying on WSDL for WS invocation
• expanding UDDI for WS discovery
• Semantic Annotations for WSDL and XML Schema W3C
  recommendation, August 2007
• OWL-S an OWL-based service ontology supplying a
  core set of markup language constructs

39
Modelling Level OWL and UML Compared

• UML structure is formally quite different from
  OWL
• Both OWL and UML are based on classes
• Key difference the notion of Property
• At first glance, the OWL property appears to be
  the same as UML association or attribute, But
• OWL
• Properties are first-class modelling elements
• UML
• An association is defined in terms of association
  ends, which must be related to a classifier
• An attribute is always within the context of a
  single class
• In OWL it is possible to state assertions on
  properties that have no equivalent in UML

40
Implementation LevelOWL and Java Compared

• Research community followed two major directions
  to express the OWL semantics using an OO language
• The definition of a meta-model that closely
  reflect the OWL syntax and semantics.
• e.g. the modelling APIs of Jena and OWL API
• The use of the Java Beans API to realize a
  complete mapping between the two models
  Kalyanpur et al.
• PropertyChecker classes to support the semantics
  of the property, axioms and restrictions
• This approach lacks an explicit meta-model

41
A Semantic Framework for MASs

• With the aim of providing a support for semantic
  Web services discovery and invocation
• WSs are supplied with a semantic description
• The discovery and invocation phases make use of
  an abstract description of the service, based on
  ontological concepts belonging to the domain
  ontologies
• Discovery semantic UDDIs are responsible of the
  semantic search
• Invocation mapping between the semantic
  description of the service and its real
  invocation
• WSDL document, from the discovery phase, is used
  in the invocation phase
• Mapping exploiting XSLT stylesheets (e.g
  MINDSWAP, OWL-S API project)

42
Semantic UDDI

• UDDI has two crucial limitations
• Its search mechanism
• WSs functionality can be described using
  classification schemes like NAICS, UNSPSC etc.
• The usage of XML to describe its data model
• Lack of explicit semantics
• But UDDI has enough support for the
  registration of semantically annotated resources
• Mapping of ontology concepts to UDDI data model
  using tModel structure
• New API to support semantic discovery of
  registered resources
• Semantic discovery algorithms

43
Semantic Search in UDDI

• Based on an externally created and operated
  matchmaker
• Semantic data are stored outside of UDDI
• A mapping of an OWL-S profile to the UDDI data
  model
• Changes to UDDI APIs for support of semantic
  search

M. Paolucci, T. Kawamura, T.R. Payne, K. Sycara,
Importing the Semantic Web in UDDI,
Proceedings of E-Services Semantic Web Workshop
(ESSW 2002), 2002
N. Srinivasan, M. Paolucci, K. Sycara, An
Efficient Algorithm for OWL-S Based Semantic
Search in UDDI Semantic Web Services and Web
Process Composition, First International
Workshop, SWSWPC 2004 96-110
M. Klusch,, B. Fries, K. Sycara, An Efficient
Algorithm for OWL-S Based Semantic Search in
UDDI AAMAS 2006
44
Business Process Management

• WSBPEL is the new standard for orchestrating
  business process using web services
• Joint IBM/Microsoft proposal, being standardized
  through OASIS
• There are some competing languages, e.g. BPML
• Supported by more platform vendors than its
  predecessors that tried to achieve similar goals,
  such as ebXML
• BPEL is supported by Microsoft, IBM, BEA, SAP,
  Hewlett- Packard, Oracle, Siebel, and others.
• Choice of process engines
• Useful in defining both concrete and abstract
  processes
• Each activity is represented as a service with a
  WSDL interface

45
Goal-Oriented Autonomic Business Process
Management

• Interesting talk on Tuesday afternoon
• Giovanni Rimassa, Birgit Burmeister
• Achieving Business Process Agility
• in Engineering Change Management
• with Agent Technology

46
Limitations of Description Logics

• Description Logics provides a snapshot of the
  World
• Describe objects in a given time frame
• They are intrinsically very static
• Difficult to represent processes (sequences of
  situations)
• There is no explicit notion of variables and
  rules
• It is impossible to say that
• X father Y and Y father Z ? X grandfather Z
• Lots of research has happened to extend RDF/OWL
  (SPARQL, SWRL,, )
• Integration of Rule-Based and Agent-Based
  Programming (e.g D4J)

47

• Concluding Remarks

48
May Agents Enhance SOA?

• Integration between local and global management
• Single agent cope with local problems
• Different agents can cooperate to cope with
  global problems
• Communication overhead
• Agent distribution
• Agent mobility
• System dynamic upgrading
• Agent evolution
• Code mobility
• Robustness and fault tolerance
• Agent replication

49
How Should an Innovative Agent-Based SOA Be?

• Enhancing interoperability through ontologies
• Enhancing security through OASIS-W3C standards
• Enable the use of workflow / rule engines

• Ontology management software are not efficient
• Security solutions for distributed systems are
  limited

50
Agents Workflow Integration

• High level means for service composition
• High level means for system programming when
  supported by visual tool
• Well known standards (i.e., xpdl ,WSBPEL, )
• Large set of software tools available

• Workflow can be used to program agent systems at
  a high abstraction level
• Workflow execution can be optimized by
  partitioning it among agents
• Agents can act as workflow engines
• System evolution can be realized updating the
  workflow to be executed

51
Agents Rule Engines Integration

• Logic separated from application code
• Logic can be dynamically updated adding/removing
  rules
• Well optimized engines

• Rules can provide agent intelligence
• Rules can be exchanged by agents for realizing
  system evolution
• Agents can provide the infrastructure for a
  distributed rule system

52
(No Transcript)
53
Team

• Agostino Poggi
• Federico Bergenti
• Lorenzo Lazzari
• Marco Mari
• Alessandro Negri
• Michele Tomaiuolo
• Paola Turci

54
Thanks and Acknowledgements

• I would like to thank
• Francesco Cirillo CEO XPLabs
• for the fruitful discussions

55
Thank you!Hope you enjoyed it!Questions?

• Paola Turci
• turci_at_ce.unipr.it

56
Selected References

• AgentLink III. Agent Technology Roadmap.
  Available from http//www.agentlink.org/roadmap/in
  dex.html
• Akkermans, H. Intelligent E-Business - From
  Technology to Value. IEEE Intelligent Systems,
  16(4)8-10, 2001.
• Bechhofer, R. Volz, and P. Lord. Cooking the
  semantic web with the OWL API. In Proc. Int
  Semantic Web Conference, pp. 659-675, Sanibel
  Island, FL, 2003.
• Beneventi, A., Poggi, A., Tomaiuolo, M., Turci,
  P. Integrating Rule and Agent-Based Programming
  to Realize Complex Systems. WSEAS Trans. on
  Information Science and Applications,
  1(1)422-427, 2004.
• Bergenti, F., Poggi, A., Tomaiuolo, M., Turci, P.
  An Ontology Support for Semantic Aware Agents. In
  Proc. Seventh International Bi-Conference
  Workshop on Agent-Oriented Information Systems
  (AOIS-2005 _at_ AAMAS), Utrecht, The Netherlands,
  2005.
• Berners-Lee, T., Hendler, J., Lassila O. The
  Semantic Web - A new form of Web content that is
  meaningful to computers will unleash a revolution
  of new possibilities. 284(5)34-43, 2001.

57
Selected References (II)

• Buhler P.A., Vidal, J.M. Towards Adaptive
  Workflow Enactment Using Multiagent Systems.
  Information Technology and Management,
  6(1)61-87, 2005.
• de Bruijn, J., Polleres, A., Lara, R., Fensel, D.
  OWL DL vs. OWL Flight Conceptual Modeling and
  Reasoning for the Semantic Web. In Proc. of the
  14th Int. World Wide Web Conference (WWW2005),
  pp. 623-632, Chiba, Japan, 2005.
• Fensel, D., Bussler, C. The Web Service Modeling
  Framework WSMF. Electronic Commerce Research and
  Applications 1(2) 113-137, 2002.
• Greenwood, D., Callisti, M. Engineering Web
  Service-Agent Integration. In IEEE Conference of
  Systems, Man and Cybernetics, 2004. Available
  from http//www.whitestein.com/resources/papers/ie
  eesmc04.pdf.
• Gibbins, N., Harris, S., Shadbolt, N. Agent-based
  semantic web services. In Proc of the 12th
  International World Wide Web Conference
  (WWW2003), Budapest, Hungary, 2003.
• Gutknecht, O., Ferber, J., Michel, F. Integrating
  tools and infrastructures for generic multi-agent
  systems. In Proc. of the 5th International
  Conference on Autonomous Agents. Montreal,
  Canada, 2001.
• Horrocks, I. and Patel-Schneider, P. F. A
  proposal for an OWL rules language. In Proc. of
  the Thirteenth International World Wide Web
  Conference (WWW 2004), pp. 723-731, 2004.

58
Selected References (III)

• Hindriks, K.V., de Boer, F.S., van der Hoek,
  W., Meyer, J.C. Control Structures of Rule-Based
  Agent Languages. In Lecture Notes In Computer
  Science, Proceedings of the 5th International
  Workshop on Intelligent Agents V, Agent Theories,
  Architectures, and Languages, Vol 1555, pp.
  381-396, 1998, London, UK Springer-Verlag.
• R. Kishore, H. Zhang R. Ramesh, Enterprise
  integration using the agent paradigm
  foundations of multi-agent-based integrative
  business information systems, Decision Support
  Systems, 42 (2006) (1), pp. 4878..
• Labrou, Y. Agents and ontologies for e-business.
  Knowledge Engineering Review, 17(1)81-85, 2002.
• Negri, A., Poggi, A., Tomaiuolo, M., Turci, P,.
  Dynamic Grid Tasks Composition and Distribution
  through Agents,. Concurrency and Computation
  Practice and Experience, 2005.
• Negri A., A. Poggi, M. Tomaiuolo, P. Turci,
  Agents for e-Business Applications, In Proc.
  AAMAS 2006, Hakodate, Japan 2006
• Papazoglou, M.P.. Agent-oriented technology in
  support of e-business. Communication of ACM,
  44(4)71-77, 2001.
• Papazoglou, M.P. The World of e-Business
  Web-Services, Workflows, and Business
  Transactions In Lecture Notes In Computer
  Science, CAiSE '02/ WES '02 Revised Papers from
  the International Workshop on Web Services,
  E-Business, and the Semantic Web, Vol 2512, pp.
  153-173, 2002. London, UK. Springer-Verlag

59
Selected References (IV)

• Silva, N., Rocha, J., Cardoso, J. E-Business
  Interoperability Through Ontology Semantic
  Mapping. In Proc. of the Processes and
  Foundations for Virtual Organizations, pp.
  315-322, 2003. Lugano, Switzerland.
• Weikum G. Special Issue on Infrastructure for
  Advanced E-services, IEEE Data Engineering,
  24(1), 2001.