Abstract
The Ubiquitous Computing envisions a world of communicating mobile devices. As mobile devices usually have limited resources, it brings up the challenge to search and invoke services at any time and any place in an accurate, efficient and adaptive way. In this paper, we propose to filter services by matchmaking the precondition of service invocation and the context information of service consumer, in order to increase the accuracy and adaptivity of service discovery and recommendation in mobile environment. We construct context ontology to describe the environment when a service consumer submits service request first. And then, we propose to formalize the precondition of a successful service invocation based on a lightweight DL theory. Finally, reasoning on context satisfiability can be used for service recommendation. A case study shows our approach works well.
Introduction
In the era of ubiquitous computing, mobile devices are expected to seamlessly discover and invoke services—ranging from public transportation schedules to real-time translation tools. However, mobile environments present unique challenges: limited screen real estate, constrained processing power, and fluctuating battery life.
Current service discovery methods often rely on simple input/output type matching. For instance, a search for a bus route might return a valid service that is no longer operational at the time of the request. This "dumb" matching fails to account for context satisfiability: the requirement that the current environment (user location, device capabilities, time) must meet the specific preconditions necessary for a service to function effectively.
Methodology: Context-Aware Filtering
The proposed framework bridges the gap between service discovery and practical utility through a three-step logical approach.
1. Context Ontology Construction
The foundation of the system is a structured context ontology. This model categorizes information into two primary domains:
- User Profile Context: Static data such as gender, specialty, and personal preferences.
- Service Invocation Context: Dynamic data including the "Purpose" of the request, the "Natural Environment" (location, date, time), and the "Computing Environment" (hardware specs like CPU/RAM and network conditions).
2. Precondition Formalization
To enable automated reasoning, the authors formalize service requirements using a lightweight version of Description Logic (DL). DL provides a formal language for knowledge representation. By utilizing a "lightweight" subset of DL, the system ensures that the reasoning process remains computationally efficient enough for real-time execution on mobile networks.
3. Reasoning Engine
The engine evaluates two critical properties:
- Executability: Determines if a specific service instance can be run by the user in their current context.
- Realizability: Checks if a service is capable of being executed under any possible context, identifying fundamentally unreachable services.
Case Study: Mobile Translation Service
The framework was validated through a detailed scenario involving a translation request. In this example, a user in Hangzhou uses an HTC G7 mobile phone to translate Chinese text into Spanish.
The reasoning engine analyzed the user's context (Request: Chinese to Spanish) against the service's supported language pairs. In this instance, the engine returned a "False" result for executability because the specific service instance only supported pairs like (English, Chinese) or (Spanish, English). By identifying this mismatch early, the system prevents the user from attempting a service invocation that would inevitably fail, thereby conserving device resources and bandwidth.
Key Insights and Results
- Multi-Dimensional Context: Effective recommendation must look beyond GPS. Hardware specifications (RAM, CPU, OS) are vital, as a service might exceed the physical capabilities of a mobile device.
- The Precondition Link: Integrating a formal logic layer for preconditions addresses a significant gap in current service standards like WSDL-S or OWL-S, which do not natively support reasoning on invocation feasibility.
- Efficiency: The choice of simplified DL operators (intersection, union, basic comparisons) makes complex reasoning viable for mobile environments.
Conclusion and Future Work
The proposed logic-based framework successfully filters services by matching real-time context against functional preconditions. While the case study demonstrates the concept's validity, future work will focus on automating the formalization process, which currently requires manual effort. Additionally, the authors plan to test the system's scalability in environments with thousands of competing services to ensure its performance in large-scale production deployments.
Summary updated