Abstract
This paper introduces a comprehensive framework that allows mobile users to access a variety of services provided by their current environment (e.g. print services). Novel to our approach is that selection and execution of services takes into account the user’s current context. Instead of being harassed by useless activities as service browsing or configuration issues, environmental services get seamlessly aligned to the user’s present task. Thus, the challenge is to develop a new service framework that fulfils these demands. The paper proposes a document-based approach; so called Context-Aware Packets (CAPs) contain context constraints and data for describing an entire service request. The core framework, Context-Aware Packets Enabling Ubiquitous Services (CAPEUS), reverts to CAPs for realising context-aware selection and execution of services.
Introduction
As we move toward ubiquitous computing, our physical environments are becoming saturated with wireless devices and services. However, the user experience often remains clunky. For example, printing a document in a new office typically requires manual discovery, compatibility checks, and configuration. The authors argue that users are currently "harassed" by these hurdles. The CAPEUS framework addresses this gap by aligning environmental services with a user’s physical context and current task automatically, moving beyond simple discovery to autonomous selection and execution.
The CAPEUS Architecture
The researchers introduced a document-based approach centered on Context-Aware Packets (CAPs). A CAP can be viewed as a highly evolved Remote Procedure Call (RPC). While a standard RPC is directed at a specific server, a CAP is "blind"—it contains data and task constraints, and the network infrastructure determines its destination based on context.
Key components of the architecture include:
- Service Access Nodes (SANs): These act as local proxies for a specific domain (e.g., a room). They evaluate incoming CAPs against local resources.
- The CAP Matcher: Residing within the SAN, this component compares CAP constraints to the actual environment. It calculates "divergence"—a measure of mismatch in location or file format—to decide whether to route the packet elsewhere or trigger a local conversion.
Context-Aware Packets (CAPs)
CAPs are XML-based documents structured into three distinct layers:
- Context Constraints: Defining the "where" and "when" requirements.
- Scripting: Allowing for complex logic and semantics (though less utilized in early prototypes).
- Data: The actual payload, such as a PDF or text file.
By treating context constraints as a "set point" (borrowing from control theory), the system constantly compares sensor data to these requirements to trigger appropriate actions.
Service Execution and Nesting
A core innovation of CAPEUS is its ability to handle interface mismatches through nesting. If a printer requires Postscript but the user provides a PDF, the SAN can autonomously wrap the original CAP inside a new one, route it to a conversion service, and then deliver the result to the target printer.
This separation of concerns allows services to remain simple and modular, as the infrastructure (the SANs) handles the complexities of "where, when, and how."
Evaluation and Results
The framework was demonstrated through two functional prototypes: Context-Aware Notes, which triggers digital messages based on physical proximity, and Task-Aligned Local Service Use, a document viewer that pre-selects contextually relevant options like the nearest printer or digital blackboard.
Quantitative observations suggest that CAPEUS significantly reduces the cognitive load on users and saves screen real estate on mobile devices by filtering menus to show only contextually valid options.
Future Work
While successful in prototype, the authors identify several areas for growth. The hierarchical routing approach (Room -> Floor -> Building) requires further testing for city-wide scalability. Additionally, future iterations aim to integrate "User Models" to automatically generate CAPs based on learned preferences, and to implement more robust security and privacy measures for context-heavy data packets.
Summary updated