A Cloud-based Approach for Context Information Provisioning

As a result of the phenomenal proliferation of modern mobile Internet-enabled devices and the widespread utilization of wireless and cellular data networks, mobile users are increasingly requiring services tailored to their current context. High-level context information is typically obtained from context services that aggregate raw context information sensed by various sensors and mobile devices. Given the massive amount of sensed data, traditional context services are lacking the necessary resources to store and process these data, as well as to disseminate high-level context information to a variety of potential context consumers. In this paper, we propose a novel framework for context information provisioning, which relies on deploying context services on the cloud and using context brokers to mediate between context consumers and context services using a publish/subscribe model. Moreover, we describe a multi-attributes decision algorithm for the selection of potential context services that can fulfill context consumers’ requests for context information. The algorithm calculates the score of each context service, per context information type, based on the quality-of-service (QoS) and quality-of-context information (QoC) requirements expressed by the context consumer. One of the benefits of the approach is that context providers can scale up and down, in terms of cloud resources they use, depending on current demand for context information. Besides, the selection algorithm allows ranking context services by matching their QoS and QoC offers against the QoS and QoC requirements of the context consumer.

💡 Research Summary

The paper addresses the growing demand for context‑aware services on mobile devices, highlighting that traditional context services cannot cope with the massive volume of sensor data generated today. To overcome storage, processing, and dissemination bottlenecks, the authors propose a cloud‑based framework that decouples context consumers from context providers through a Context Broker employing a publish/subscribe (Pub/Sub) model.

The framework consists of three main components: (1) Context‑aware Web Services (CAWS) that act as consumers, expressing their interest in specific context topics (e.g., location, temperature, user activity) and providing QoS (Quality‑of‑Service) and QoC (Quality‑of‑Context) requirements; (2) Context Brokers, which manage subscriptions, receive context updates from providers, and select the most suitable providers based on a multi‑attribute decision algorithm; and (3) Cloud‑based Context Services, deployed as SaaS, which aggregate raw sensor data, compute high‑level context, and expose it via web service interfaces.

A central contribution is the definition of QoC indicators—precision, freshness, temporal resolution, spatial resolution, and probability of correctness—as well as traditional QoS metrics such as latency, availability, and cost. Each consumer can assign weights and minimum thresholds to these attributes, thereby expressing a nuanced quality profile for each required context type.

The multi‑attribute selection algorithm computes a score for every candidate context service per topic. The score is a weighted sum of the service’s QoC and QoS values, normalized against the consumer’s requirements. Services are then ranked, and the broker forwards the top‑ranked (or top‑N) services to the subscriber. The algorithm works both in a single‑cloud scenario and across multiple clouds, allowing providers to scale resources up or down according to demand.

The authors argue that placing context services in the cloud yields several benefits: (i) elastic scalability—providers can provision additional compute and storage resources on demand; (ii) economies of scale—shared cloud infrastructure reduces operational costs; (iii) pay‑as‑you‑go pricing—consumers can select services based on cost and quality; and (iv) simplified integration—CAWS need not manage context acquisition or storage themselves.

Related work is surveyed, showing that prior context‑aware middleware often suffers from limited scalability, poor interoperability, and lack of quality guarantees. The proposed approach differentiates itself by leveraging cloud elasticity and by explicitly incorporating QoC into the service selection process.

The paper also discusses open challenges: security and privacy of context data, handling of topic collisions, consistency across multiple cloud providers, and the computational overhead of the selection algorithm. The authors acknowledge the absence of experimental validation and suggest future work involving large‑scale simulations, performance benchmarking, and extensions to domains such as smart cities and healthcare IoT.

In conclusion, the study presents a novel architectural pattern that combines cloud deployment with a broker‑mediated Pub/Sub mechanism and a quantitative QoC/QoS matching algorithm. This combination promises to deliver high‑quality, context‑aware services to mobile users in a scalable, cost‑effective manner, while opening avenues for further research on optimization, security, and real‑world deployment.