Walid-Khaled Hidouci

A real-time feedback scheduler for environmental energy with discrete voltage/frequency modes

This paper, addresses the problem of real-time scheduling in ambient energy harvesting systems with discrete voltage/frequency modes through the use of a feedback scheduler. The main idea explored in this paper is to reduce the processor speed in proportion to the amount of available energy and processor load. The evaluation of this new real-time feedback scheduler shows experimentally a good compromise between the available energy and the processor workload, and allows a progressive selection of processor speeds. In this work, we compare several battery thresholds knowing that the processor performs the tasks at maximum speed when the battery storage is over the thresholds.

A Novel BILP Model for Energy Optimization Under Data Precision Constraints in Wireless Sensor Networks

In this letter, we propose a binary integer linear programming (BILP) model of the problem of Energy Minimization under the constraint of Data Precision in the context of correlated data collection in wireless sensor networks, called EMDP. The exact solution of our BILP model determines, in each round of data collection, the role of each node in terms of sensing, data relaying, and processing. It gives the baseline for optimal network operations and helps characterizing the complexity of EMDP problem. Moreover, we propose a heuristic solution, namely, CORAD, which is an energy-aware correlation-based adaptive dynamic clustering algorithm for data collection.

From KDD to KUBD: Big Data Characteristics Within the KDD Process Steps

Big Data is the current challenge for the computing field not only because of the volume of data involved but also for the amazing promises to analyze and interpret massive data to generate useful and strategic knowledge in various fields such as security, sales and education. However, the massive volume of data in addition to other characteristics of Big Data such as the variety, velocity, and variability require a whole new set of techniques and technologies, which are not yet available, to effectively extract the desired knowledge. The KDD (Knowledge Discovery in Databases) process has achieved excellent results in the classical database context and that is why we examine the possibility of adapting it to the Big Data context to take advantage of its strong and effective data processing techniques. We introduce therefore a new process KUBD (Knowledge Unveiling in Big Data) inspired from the KDD process and adapted to the Big Data context.

RTSBL: Reduce Task Scheduling Based on the Load Balancing and the Data Locality in Hadoop

We address load balancing and data locality problems in Hadoop. These two problems limit its performance, especially, during a reduce phase where the partitioning function assigns the keys to the reducers based on a hash function. We propose in this paper a new approach to assign the keys based on the reducers’ processing capability in order to ensure a good load balancing. In addition, our proposed approach called RTSBL takes into consideration the data locality during the partition. Our experiments prove that RTSBL achieves to up 87% improvements in the load balancing and 3\(\times \) improvements of the data locality during the reduce phase in the standard Hadoop.

Activities scheduling algorithms based on probabilistic coverage models for wireless sensor networks

Area monitoring using Internet and barrier coverage is a typical application of wireless sensor networks. The main concerns in this type of applications are coverage efficiency and sensor energy conservation. For that, many activities scheduling algorithms are proposed in the literature. Unlike prior efforts based on an unrealistic binary sensor coverage model, this paper proposes three efficient activities scheduling algorithms based on realistic sensor coverage models. The first algorithm (C1L-PBC) is centralized and it is based on a coverage graph. The second algorithm (D1L-PBC) is distributed and it ensures 1-barrier coverage; whereas, the third one (D2L-PBC) is also distributed and it guarantees 2-barrier coverage. The obtained experimental results show that the proposed algorithms can effectively guarantee the barrier coverage and prolong the sensor network lifetime.

A real-time feedback scheduler based on control error for environmental energy harvesting systems

This paper addresses a real-time scheduling problem inherent to energy harvesting real-time systems. Traditionally, the energy saving problem is solved mainly by taking into account the tasks scheduling parameters such as worst-case execution time and period. In this work, we construct a feedback control scheduling scheme in which a discrete processor speed is assigned according to the control error and available energy. The real-time control tasks would get high processor speeds when their control errors increase. The experimental evaluation of this solution verifies that the feedback scheduling system based on control error gives a good compromise between available energy and systems performance.

A Real-Time Feedback Scheduler for Environmental Energy with Discrete Voltage/Frequency Modes

This paper, addresses the problem of real-time scheduling in ambient energy harvesting systems with discrete voltage/frequency modes through the use of a feedback scheduler. The main idea explored in this paper is to reduce the processor speed in proportion to the amount of available energy and processor load. The evaluation of this new real-time feedback scheduler shows experimentally a good compromise between the available energy and the processor workload, and allows a progressive selection of processor speeds. In this work, we compare several battery thresholds knowing that the processor performs the tasks at maximum speed when the battery storage is over the thresholds.

SISO modelling of a web server to be controlled by a Feedback control scheme

Performance modeling is an important topic in overload control for web servers. Several attempts have been made to create performance models for web servers. The paper describes modeling a Web server to be controlled by Feedback control scheme. Feedback command theory was initially used to control of industrial processes. Its use for the control of performance software is recent. It provides a number of mathematical tools which can be used to analyze the stability of the commanded system and find the best adjustment that responds to the performance criteria. Our approach proceeds in two steps: system identification and controller design. In system identification, we construct mathematical models of the target system in forms of discrete transfer function focused on single-input single-output (SISO) systems. The role of controller is to modify the transfer function of the target system with regard to the control error between the reference value and the output value