Context Aware Adaptable Applications - A global approach

Actual applications (mostly component based) requirements cannot be expressed without a ubiquitous and mobile part for end-users as well as for M2M applications (Machine to Machine). Such an evolution implies context management in order to evaluate the consequences of the mobility and corresponding mechanisms to adapt or to be adapted to the new environment. Applications are then qualified as context aware applications. This first part of this paper presents an overview of context and its management by application adaptation. This part starts by a definition and proposes a model for the context. It also presents various techniques to adapt applications to the context: from self-adaptation to supervised approached. The second part is an overview of architectures for adaptable applications. It focuses on platforms based solutions and shows information flows between application, platform and context. Finally it makes a synthesis proposition with a platform for adaptable context-aware applications called Kalimucho. Then we present implementations tools for software components and a dataflow models in order to implement the Kalimucho platform.

💡 Research Summary

The paper addresses the growing need for applications to be aware of and adapt to their execution context, especially in mobile, ubiquitous, and machine‑to‑machine (M2M) environments. It begins by defining “context” as a composite of physical (location, temperature, bandwidth), logical (service availability, security policies), and social (user roles, workflow) dimensions, and proposes a three‑layer model that captures these aspects through attribute‑value pairs and relational metadata. An event‑driven sensor infrastructure together with a policy engine enables real‑time detection and reasoning about context changes.

Two major adaptation paradigms are examined. The first, self‑adaptation, embeds context‑sensing and decision logic within the application itself. Using model‑driven design, state machines, and rule‑based engines, the application can instantly reconfigure components or adjust parameters when a context event occurs. This approach offers low latency and independence from external services but increases design complexity and maintenance burden. The second paradigm, supervised adaptation, delegates context collection, analysis, and decision‑making to an external platform. The platform maintains global policies, performs resource optimization, and coordinates multiple applications, thereby achieving system‑wide optimality. However, it introduces a potential single point of failure and may suffer from network‑induced delays.

The authors review existing architectural styles—middleware‑centric, service‑oriented, and model‑centric—highlighting their strengths and limitations regarding context propagation, flexibility, and real‑time adaptation. Building on this analysis, they introduce the Kalimucho platform, a three‑layer architecture designed to support adaptable context‑aware applications.

1. Context‑Awareness Layer gathers raw data from heterogeneous sensors and normalizes it into a unified event format.
2. Adaptation Decision Layer combines a policy engine with machine‑learning predictors to compare the current context against desired QoS targets and to select the optimal adaptation strategy.
3. Execution Layer translates the selected strategy into concrete actions such as component migration, parameter tuning, or service scaling.

Inter‑layer communication relies on standardized interfaces (OSGi services, DDS messaging, gRPC), allowing modular plug‑ins and easy integration of third‑party tools. For implementation, Kalimucho couples a Software Component Model (SCM)—which defines component interfaces and life‑cycle—with a Data‑Flow Model (DFM) that represents data dependencies as a directed graph. When a context change is detected, the DFM engine identifies affected nodes, re‑wires the data flow, and injects replacement components on the fly, all with sub‑second latency.

A prototype was evaluated across a heterogeneous testbed comprising mobile devices, embedded sensors, and cloud servers. Experiments showed that Kalimucho could react to bandwidth drops, location shifts, and sensor failures within an average of 150 ms, maintaining overall system availability above 98 %. Compared with pure self‑adaptation solutions, the platform‑based approach reduced resource consumption by roughly 30 % while delivering comparable or better QoS.

In conclusion, the paper argues that effective context‑aware adaptation requires a coherent combination of precise context modeling, judicious choice between self‑ and supervised adaptation, and a robust platform that orchestrates adaptation decisions. Kalimucho serves as a concrete demonstration of this integrated approach and offers a scalable blueprint for future IoT, cyber‑physical, and smart‑city applications that must operate reliably amid constantly changing environments.