Stella Kafetzoglou

Energy-efficient framework for data gathering in wireless sensor networks via the combination of sleeping MAC and data aggregation strategies

In this paper, we introduce and evaluate a data collection and transmission paradigm in Wireless Sensor Networks (WSNs), where the primary objective is energy efficiency, while at the same time it is flexible and adaptive to support various Quality of Service (QoS) application constraints. The proposed framework combines two different energy-saving methods, one at the application layer, namely data aggregation, and one at the Medium Access Control (MAC) layer by appropriately adopting sleeping mechanisms. This is achieved by following a cross-layer approach where local information about the status of each sensor, including its sleep/awake schedules, feeds a subsequent data aggregation phase, which gathers correlated data and aggregates it to a single packet towards its way to the sink by applying a fully localised, distributed, probabilistic method. Simulation results demonstrate the significant energy savings achieved by our proposed framework, especially when compared to conventional MAC layer approaches and/or data aggregation approaches.

The ArguGRID Platform: An Overview

The ArguGRID project aims at supporting service selection and composition in distributed environments, including the Grid and Service-oriented architectures, by means of argumentative agents, an agent environment, a service-composition environment, Peer-to-Peer technology and Grid middleware. Agents are argumentative in that they use argumentation-based decision-making and argumentation-supported negotiation of services and contracts. The integration of all technologies gives rise to the overall ArguGRID platform. In this paper we outline the main components and the overall functionalities of the ARGUGRID platform.

Sensor Network Grids: Agent Environments Combined with QoS in Wireless Sensor Networks

We present a distributed systems architecture that uses Grid computing to combine basic nodes of wireless sensor networks with complex sensor nodes of wired networks. Three kinds of complex sensor nodes are identified: objects, agents and containers. The decision making capabilities of the complex nodes are then combined with a quality of service (QoS) framework for gathering data from the wireless sensor nodes. The resulting combination provides a powerful conceptual framework for developing ambient intelligence and ubiquitous computing applications on the grid.

An Efficient QoS-Constrained Data Aggregation and Processing Approach in Distributed Wireless Sensor Networks

In this paper an efficient Quality of Service (QoS)- constrained data aggregation and processing approach for distributed wireless sensor networks is introduced and evaluated. The objective of the proposed approach is to aggregate data on the fly at intermediate sensor nodes in order to improve the operational efficiency and effectiveness of the sensor networks, while at the same time still satisfying the latency and measurement quality constraints. One of the key features of the proposed approach is that the task QoS requirements are taken into account to determine when and where to perform the aggregation in a distributed fashion. The performance of the proposed approach is analyzed and evaluated, through modeling and simulation, under different data aggregation scenarios and traffic loads. The impact of several design parameters and tradeoffs on various critical network and application related performance metrics, such as the energy efficiency and end-to-end latency, are also evaluated and discussed.

On the modeling of data aggregation and report delivery in QoS-constrained sensor networks

In this paper the problem of efficient data aggregation and report delivery in QoS-constrained distributed sensor networks is investigated. An analytical statistical model is introduced, to represent the data aggregation and report delivery process in sensor networks, with specific delivery quality requirements in terms of the achievable end-to-end delay and the successful report delivery probability. Furthermore, a lower bound approximation of the corresponding successful report delivery probability is obtained. The tradeoffs among various design parameters, as well as the impact of these parameters on the successful report delivery probability and its corresponding bound, are also discussed and evaluated

A holistic approach for personalization, relevance feedback & recommendation in enriched multimedia content

As the prerequisites of production houses, broadcasters, advertising agencies and online publishing companies for enriched multimedia content increase rapidly, the need of innovative methods for the effective creation of enriched multimedia content is undeniable. Stemming from this need, in this paper we focus on the design, development and evaluation of a framework consisting of personalization, relevance feedback and recommendation mechanisms, as a principal method for the creation of enriched multimedia content targeted to each user’s needs, preferences and interests. As the multimedia content proliferates along with its consumption by the users, more effective ways of presenting it to the viewers are demanded in order to facilitate them with the multimedia content search and selection and improve their Quality of Experience (QoE). The main contribution of the paper is the introduction of a holistic framework that offers personalized enriched multimedia content, by extending the recommendation process to the set of enrichments that accompany the video except from the video itself and by collecting explicit and implicit relevance feedback from the interactions of the user with both the video and its enrichments. We evaluate the proposed framework following a two-step approach. Firstly, we perform extended experiments by applying reasonably simulated user interactions, in order to calibrate its parameters that refer to multiple aspects of the enriched multimedia content, aiming at high performance in terms of QoE. Here, most importantly, we have shown that appropriately designing the enrichments and considering users’ interactions with them allows for achieving a better quality in inferring users’ profiles in many realistic cases. Secondly, we integrated our proposed recommender framework within the MECANEX streaming platform in order to perform user studies about its usability within a realistic environment of use.

On the Optimization of a Probabilistic Data Aggregation Framework for Energy Efficiency in Wireless Sensor Networks

Among the key aspects of the Internet of Things (IoT) is the integration of heterogeneous sensors in a distributed system that performs actions on the physical world based on environmental information gathered by sensors and application-related constraints and requirements. Numerous applications of Wireless Sensor Networks (WSNs) have appeared in various fields, from environmental monitoring, to tactical fields, and healthcare at home, promising to change our quality of life and facilitating the vision of sensor network enabled smart cities. Given the enormous requirements that emerge in such a setting—both in terms of data and energy—data aggregation appears as a key element in reducing the amount of traffic in wireless sensor networks and achieving energy conservation. Probabilistic frameworks have been introduced as operational efficient and performance effective solutions for data aggregation in distributed sensor networks. In this work, we introduce an overall optimization approach that improves and complements such frameworks towards identifying the optimal probability for a node to aggregate packets as well as the optimal aggregation period that a node should wait for performing aggregation, so as to minimize the overall energy consumption, while satisfying certain imposed delay constraints. Primal dual decomposition is employed to solve the corresponding optimization problem while simulation results demonstrate the operational efficiency of the proposed approach under different traffic and topology scenarios.

Performance Evaluation of a Robust Data Aggregation Approach in Diverse Sensor Networking Environments

In wireless sensor networks, a sensor node, besides being able to act on the environment, is also a network entity that performs networking-related functionalities and could be integrated within the networking and grid environment. In this chapter, we adopt a probabilistic approach that presents a robust and generalized framework for performing in-network processing in order to improve the overall data collection and monitoring process in a sensor network. The effectiveness and applicability of the framework under investigation in a diverse set of networking environments that present different constraints, limitations, and physical characteristics are demonstrated via modeling and simulation, under different traffic loads. Furthermore, different ways of dynamically varying and configuring parameters related to the aggregation process are investigated and evaluated for different applications and networking environments (terrestrial and underwater sensor networks).

Reputation-Based Trust in federated testbeds utilizing user experience

In this paper a novel reputation-based trust framework in federated testbeds is presented, termed as Federated Trust and User Experience (FTUE) Framework. Our aim is to assist peer experimenters in properly selecting resources from the federated community by constructing a quantitative view of the trustworthiness of each testbed. The FTUE framework combines the perceived user experience, as provided by user feedback, with the monitoring data retrieved for the respective experiments. Several situations that may arise in conducting and evaluating experiments were identified, properly formulated and simulated. The results presented demonstrate that the FTUE framework has the ability to dynamically adjust the reputation under challenging conditions, as well as being resilient to malicious attacks. Therefore, it could provide a paradigm for dynamic reputation building and trustworthiness among testbed services and users.

Discovery of Resources in a Distributed Grid Environment Based on Specific Service Level Agreements (SLAs)

In distributed environments it is necessary to provide tools that describe and guarantee the level of Quality of Service of the provided services. These guarantees can be specified in terms of Service Level Agreements (SLAs), which in turn can be used for the management and monitoring of Quality of Service attributes. The provided services supply the appropriate interfaces through which the service providers can allocate and manage their resources. The emphasis of this work is placed on the discovery of resources in a distributed environment, based on Quality of Service attributes, which can be expressed in terms of SLAs between the service provider and the client. For this purpose a registry that supports dynamic registration and contextualized discovery of resources is used. GRIA software platform has been adopted as the basis for the provisioning of the resources to be managed and the design of the proposed SLA-based resource discovery mechanism, as it provides a service-oriented infrastructure, designed to support service provisioning in a secure, interoperable, and flexible manner.