Chapter 10: Execution Models

1

• Chapter 10Execution Models

Service-Oriented Computing Semantics, Processes,
Agents Munindar P. Singh and Michael N. Huhns,
Wiley, 2005
2
Highlights of this Chapter

• Interoperation Architecture
• Messaging
• Peer-to-Peer Computing
• Enterprise Service Bus
• CORBA
• Jini
• Grid Computing

3
Interoperation (Standards and Technologies)

• Requires surmounting a series of challenges
• Transport HTTP, SMTP, SIP
• Messaging XML (including XQuery ), SOAP
• Data and structure WSDL
• Finding and binding UDDI, QoS techniques
• Semantics ontologies (RDF, OWL, IEEE SUO, Cyc)
• Transactions WS-Coordination, WS-AtomicTransactio
  n, WS-BusinessActivity
• Process OWL-S, WS-CDL, PSL, BPEL4WS
• Policy XACML
• Dynamism FIPA AMS, RuleML, Jason
• Cooperation FIPA ACL, multiagent systems

4
Application Interoperation
5
Architectural Elements

• Low-level (included in app server)
• Directories, messaging
• Data and process interoperation
• Metadata and transformations
• Routing
• Rules engine
• Business process
• Modeling and execution engine

6
Invocation-Based Adapters

• Common in distributed object settings (EJBs,
  DCOM, CORBA)
• Synchronous blocking method invocation
• Asynchronous nonblocking (one-way) method
  invocation with callbacks
• Deferred synchronous (in CORBA) sender proceeds
  independently of the receiver, but only up to a
  point
• Execution is best effort, at most once
• More than once is OK for idempotent operations,
  not otherwise

7
Message-Oriented Middleware 1

• Analogous to store and forward networks
• Queues point to point
• Support posting and reading messages
• Topics logical multicasts
• Support publishing and subscribing to
  application-specific topics
• Thus more flexible than queues
• Some messages correspond to event notifications

8
Message-Oriented Middleware 2

• Inherently, asynchronous
• Supports loose coupling
• Reliability cannot guarantee successful message
  delivery, but can provide failure notification
• Usually used through an invocation-based
  interface
• By polling or via registered callbacks
• onMessage() method of message-driven beans

9
Enterprise Service Bus

• An abstraction separating enterprise services and
  transport
• Supports services as units of functionality
• Supports routing of messages (via MoM or Web
  services or anything else)
• Enables an architectural style in which
  transformers convert message formats
• Often accompanied with modules for process,
  policy, logging, identity management,

10
Peer-to-Peer Computing

• Models of computation
• Symmetric client-server Each party can query
  the other, thereby giving each power over the
  other at different times
• Doesn't fundamentally look beyond client-server
• Asynchrony While the request-response paradigm
  corresponds to pull, asynchronous communication
  corresponds to push
• Undesirable push applications that place their
  entire intelligence on the server (pushing) side
• Federation of equals When the participants can
  enact whatever protocols they see fit

11
Chapter 10 Summary

• Increasingly, interoperation architectures
  promote loose coupling and arms-length
  relationships
• Hence focus on messaging
• Similar challenges have been addressed multiple
  times but service improvements result in easier
  composition and deployment of services