Seminar: Trends in Service-Oriented Computing - Model-driven Integration

58315302
4
Networking and Services
Advanced studies
Year Semester Date Period Language In charge
2015 autumn 01.09-08.12. 1-2 English Lea Kutvonen

Lectures

Time Room Lecturer Date
Tue 14-16 C220 Lea Kutvonen 01.09.2015-29.09.2015
Tue 14-16 C220 Lea Kutvonen 03.11.2015-03.11.2015
Tue 14-16 C220 Lea Kutvonen 17.11.2015-08.12.2015

Information for international students

NEWS!

  • Having received a few excuses on being forced to miss the first session on this seminar, I find it more sensible to postpone the starting session by a week. We meet first time on Tuesday 8.9!
  • Note Schedule page - that is the one that is kept updated regularly.

Service oriented computing

Service oriented architecture (SOA) refers to a commonly known architecture pattern where clients and servers are loosely coupled and find each other at runtime based on the service type required, instead of the server identity. Service oriented computing is an extension of that paradigm and brings in business process management to specify how the clients and servers actually communicate with each other.

Service integration and interoperability

The terms service integration and  service interoperability are more or less synonyms in research literature. In both cases, the aim is to create a temporary or long-term composition or collaboration of services that potentially are provided by different organisations (governmental organisations, private companies, groups). Integration refers to the need of enforcing that composite service to perform a long-term activity with a clear goal, just as if it were a single system. Interoperability term is used when the autonomy of the involved organsations is emphasised. Dependind on the literature, one or the other is used. Other related terms include enterprise interoperability, refering to the need of organisations to get work done together, mostly with the same technical solutions as in service interoperability.

Basically, the aims fall into a few simple goals:

  • interoperability (or integration) refers to the state of affairs under which the partners (services) are able to meaningfully conduct collaborative work together; this involves capability of exchanging messages for knowledge exchange, for proposals and commitments, establishing trust, for example. 
  • interoperability is layered into technical, semantic and pragmatic interoperability;
    • technical interoperability covers physical means to transport messages,
    • semantic interoperability covers similar interpretation capabilities for the data exchanged in messages, operation names, and service type related details, while
    • pragmatic interoperability covers similar enough understanding of joint business process and private decisions on willingness to participate such processes inlucing trust and privacy needs.

Models in service engineering and operation

A model is an abstract view to a system, a piece of information, or an activity. In the software engineering process, models are used for abstractly describing the plans, and sometimes, formally specifying the systems in a way, that analysis on the properties of the system can be done.

However, in model-driven engineering and operation, the models have more specific meaning. In that context, a model is a formal abstraction that is structured in a way that it can be used for generating or controlling the actual system (potentially in several refining steps). In model-driven engineering, the ability to preserve the reverse engineering relationships from the created system to the more abstract models so that changes at any level can be flooded to all other abstraction levels too.

In model-driven service operation, the models are used to control the runtime behaviour of the service. In failure situations, the system is able to report on the failure, or event to do some self-recovery actions. The capability to self-recover and self-adapt based on available context information is of essential interest.

Model-driven integration of services in inter-enterprise environments

In this seminar, we study projects that have published original research papers on model-driven integration of services and compare the solutions. We focus especially on situations where integration is done across organisational boundaries. This causes involvement of supporting utilities, such as ontologies, model transformations and contracting.

Background literature