Integration Issues of an Ontology Based Context Modelling Approach

Thomas Strang, Claudia Linnhoff-Popien, Korbinian FrankICWI 2003

Abstract

In this paper we analyse the applicability of our ontology based context modelling approach, considering a range of use cases. After wrapping up the model and the Context Ontology Language (CoOL) derived from it, we introduce some interesting applications of the language, based on a scenario showing the challenges in context aware service interactions. We focus on two submodels of our model for context aware service interactions, namely Context Bindings and Context Obligations, and demonstrate how to integrate them into different existing service architectures.

Introduction

In the evolving landscape of ubiquitous and pervasive computing, services must understand a user's "context"—such as location, time, or device constraints—to provide meaningful value. A printing service, for instance, is only useful if it recognizes the user's current location and the operating hours of nearby facilities.

The primary hurdle is interoperability. Disparate services may use different units (e.g., kilometers vs. miles) or coordinate systems (e.g., GPS vs. street addresses). Without a formal, machine-readable method to describe these contextual facts and their interrelationships, seamless collaboration remains elusive. This paper addresses this gap by proposing a standardized context language integrated into existing web service architectures.

The ASC Model and CoOL

The authors introduce a multi-layered framework for context modeling:

  1. The ASC Model: The core conceptual framework consisting of Aspects (general dimensions like "Spatial Distance"), Scales (specific measurement units like "Kilometers"), and Context Information (the actual data).
  2. Operations: To facilitate translation between scales, the model employs IntraOperations (e.g., km to miles) and InterOperations (e.g., combining distance and time to derive speed). MetricOperations allow for the comparison and sorting of values.
  3. CoOL (Context Ontology Language): A concrete implementation of the ASC model using Semantic Web languages (DAML+OIL or OWL). This enables validation and processing via reasoners like OntoBroker using F-Logic.

Integration and Middleware Mediation

A key contribution of this work is the formalization of how context is integrated into service interactions. This is achieved through Context Bindings, which link service parameters to specific context aspects, and Context Obligations, which define service constraints (e.g., operating radius).

To manage these interactions, the authors suggest a middleware-based approach using an Intermediate SOAP Node, referred to as a Context Management Access Point (CMAP). This node can intercept service calls to resolve contextual data—such as looking up a user's GPS coordinates based on a phone number—before the request reaches the final service provider.

Delegated Service Invocation via an Intermediate Node

Extending Semantic Web Services

The paper proposes extending the DAML-S service ontology by adding a fourth pillar: ServiceContext. This new layer complements the existing Profile, Model, and Grounding components, allowing context to be treated as a first-class citizen within the service description.

DAML-S with Context Extension

By defining IntraOperations within this framework, systems can automatically convert data between measurement systems, resolving a significant interoperability challenge in distributed environments.

Results and Future Work

The research demonstrates the feasibility of extending standard protocols like WSDL with XML attributes to flag context-bound parameters. Furthermore, it shows how SOAP headers can carry "in-band" context identifiers to facilitate real-time resolution by context providers. Notably, the authors observe that Quality of Service (QoS) parameters can be managed as specialized forms of context within the ASC framework.

Future work will focus on the stability of emerging standards like DAML-S, expanding the catalog of pre-defined ASC artifacts (e.g., time, geography, weather), and addressing the reasoning overhead introduced by F-Logic in high-speed, real-time environments.

Summary updated

BibTeX
@inproceedings{strang2003integration,
  author = {Thomas Strang and Claudia Linnhoff-Popien and Korbinian Frank},
  title = {{Integration Issues of an Ontology Based Context Modelling Approach}},
  booktitle = {ICWI},
  year = {2003},
}
Figure