Hicham Medromi’s research while affiliated with University of Hassan II Casablanca and other places

Containers are a lightweight cloud infrastructure deployment technology that has gained unprecedented popularity due to their flexible portability, low overhead, and resources elasticity. Leveraging containerization as an alternative to virtualization, various applications dependencies may be deployed in a self-contained piece with separated binary and library files. Hence, this alternative has been suggested as a suitable option for more interoperable application packaging in the edge, fog, or cloud systems. With the rising focus on container-based virtualization, the need to investigate the implementation and orchestration of containerized clusters has emerged as a key research issue. To fill this gap, this paper seeks to systematically consolidate the existing literature following the trend of three research areas observed over years. The first research area englobes an overview on the key prerequisites that a standard federated containerized architecture must meet, followed by some prominent containers use cases. The second research area focuses then on a performance comparison of various container solutions with regard to hypervisor technologies and bare-metal deployment. Eventually, containers infrastructure scheduling techniques are discussed as a third research area in the form of a granular insight into resource allocation and auto-scaling policies.

As cloud traffic never stops growing in minutes, hours and daily basis, proactive cloud resources orchestration becomes a prerequisite. In this paper, we investigate an Amazon case study, in which we intend to compare respectively univariate and multivariate predictions of multimodal AWS instances demand and their related instances resources occupancies. For this purpose, we implemented four nonlinear deep neural network (DNNs) models, namely: LSTM, GRU with their bidirectional variants BiLSTM and BiGRU. Experimentation test scenarios demonstrated the performance of BiGRU models above other candidate models, achieving until (0.71, 0.11, 0.26, 0.97) of RMSE values, respectively while predicting four instances families’ future demands. Adopting an extended BiGRU version, we further demonstrate how multivariate predictions remain much less accurate than univariate forecasting scenarios.

As the combination of FW(Fixed-Wing) and VTOL(Vertical-Taking-Off-and-Landing) UAV (Unmanned Aerial Vehicle), the hybrid drone is more accepted as its multipurpose application. With the aid of CFD method for test and RSM (Response Surface Methodology) for optimization, this article presents new concept of canard wings that were integrated to the fuselage of the UAV. It combines the conventional delta wing with winglet and the canard configuration. Based on the requirements and the limitations of the design, the lift was optimized and distributed respectively to the main wing (90-95%) and to the front wing (5-10%). Multi-level optimization approach using RSM was developed for the optimization. At the first level, we optimize the delta wing, then a second level is applied to evaluate the final design of the canard wing from the previous step. By comparison with the initial design, this concept provides enough lift with less drag in cruise model. Besides, the control surface of the canard wing in the front could replace the tail wing that was used for trimming and dynamic controls which turns the drone into a tailless vehicle which reduce also the weight of the UAV.

Interest in electric unmanned aerial vehicles (UAVs) has grown rapidly in recent years, and their applications have expanded and diversified considerably since they first appeared, for both commercial and private purposes. Thanks to their ability to perform challenging and hazardous tasks with high mobility, safety, and low cost. As academic researchers, we are concerned with commercial multi-rotor UAVs, which are revolutionizing many public services, including search and rescue operations, wireless coverage, delivery services, precision agriculture, wildlife surveys, and real-time surveillance. One of the UAVs main issues when it comes to mobility is the limited energy autonomy/endurance. Many types of power supplies can be implemented in UAVs, each with its specific strengths and shortfalls in terms of size, charging/discharging time, energy density and power density. This paper focuses on UAVs energy aspect, with a comprehensive review of the main power sources available for multi-rotors UAVs, and energy management systems to uncover gaps and provide further insights and guidelines for future research.

The sharp increase of multimodal cloud resources demand makes it more challenging to design rightsized virtual instances. Inefficient embedding of high sized instances into the substrate resource network has led to numerous resource underutilization issues, which further constitute a key driver to repetitive reallocations of virtual instances. Besides, repetitive reconfigurations of virtual network instances go through a process of intra- or inter-cloud migration that provokes additional increase in power consumption. This paper proposes to solve these mutual challenges through a proactive, power efficient and hybrid Virtual Network Embedding (VNE) approach. We first formulated a Mixed Integer Linear Programming (MILP) model purposing to maximize total power efficiency of intra Data Center (DC) and inter networking resources as a function of EC2 instances requests rates. Leveraging the AWS cloud as a primary case study for this paper, the suggested VNE combines a multi-stage hybrid Virtual Node Embedding (VNoE) policy with an adaptive multistep consolidated Virtual Link Embedding (VLiE). As a starting point, a Green-Location aware - Global Topology Ranking (GLA-GTR) is designed as a primary ranking process suggesting the greenest substrate DCs locations with their related delivery networks. After implementing our proposal on a real AWS backbone network topology, simulation results indicated the efficiency of the proposed VNE approach. The Stacked Denoising Auto Encoders - Bidirectional Gated Recurrent Unit - Resources Vector Matching VNoE (SDAE-BiGRU-RVM VNoE) policy achieved a power decrease of 14.61%, 14.95% and 17.21% compared to BiGRU-RVM-VNoE, BiGRU-BF-VNoE and BF-VNoE policies, respectively. Accordingly, the suggested policy has reached significant power efficiency and overall maximized resource utilization.

Cloud-fog based industries are entailing today greedy energy costs, given the wide multiplication of their AI models and distributed BD frameworks implementations. This paper conducts a newsworthy inquiry purposing to study at which extent IT community are conscious about energy evaluation of their hardware and software implementations, and whether they are in line with sustainable implications toward efficient AI and BD deployments. Through an analysis of responses to the first inquiry questions, we were able to address the residing interoperability between AI models, distributed BD frameworks, and cloud-fog systems. Unfortunately, only 10% of respondents were adopting energy metrics when evaluating their implementations. Even worse, multi-level energy consumption measurement techniques were not evident to most respondents. Accordingly, we provided a useful guideline of various multi-level energy and power estimation approaches. Hereafter, we devoted in accordance with both inquiry and literature results some essential parts for analyzing emerging efficient DNNs and distributed BD implementations. These endeavors were mainly manifested in the form of efficient reconfigurable accelerators designs based on Processing-In-Memory and Processing-Near-Memory architectures. To serve eventually IT community with other tangible solutions, we proposed two roadmaps opening up to the possibility of investigating sustainable actions covering hardware, software, and data levels.

City transportation has three basic components that create the essential environment for its functioning and the social welfare namely infrastructure, operational assets, and management policies. The key focus of this article is on understanding long-term distribution of transport demand in order to build bundling networks. To achieve this aim, we provide a hybrid machine-learning approach using a combination of several clustering and forecasting algorithms that are considered efficient given the key performance indicators obtained. This approach involves combining two types of algorithms: clustering and prediction algorithms. Based on simulated benchmarks, results indicated that the clustering phase is still appropriate using the k-means algorithm. To improve the k-means results, we measured 30 validation indices to estimate the number of clusters. In so doing, not only does it want to validate the clusters but also to identify the optimal k. To evaluate forecast accuracy in the demand prediction phase, we used the standard key performance indicators, namely MSE, RMSE, MAPE and R². The SVM algorithm has been judged as the most efficient prediction algorithm based on average values of the obtained metrics.

In recent years the Unmanned Aerial Vehicles (UAVs) are increasingly adopted in various applications of the aerial transportation, meanwhile both the two main classifications, the fixed wing and multi-rotors drones, are limited by their own drawbacks in different specialized domains. In order to improve the short duration of multi-rotor drone and the strict conditions for taking-off and landing of the fixed-wing, an innovative hybrid UAV with integral advantages of "delta+canard wing" and vertical lift-fans has been developed. Meanwhile, the parametrial optimization with multi-objects are discussed based on surrogate model to determine the conception of the delta wing and canard wing, and only the design speed of cruise model is considered, thus excluding the lift fan. Firstly, several parameters are chosen according to precedent experiences and researches, together with the optimal objects and constraint conditions determined by design conditions which are concluded into the typical multi-objective optimal problem; secondly, with the aid of computational fluid dynamics (CFD) limited design points generated by meta-model will make up the design of experiment (DOE) to create the response surfaces by means of Kriging method; lastly, the adaptation of the multi-objective genetic algorithm(MOGA) integrating in modeFRONTIER helps to reach the heuristic approach optimization for this problem. Therefore, the final conception of the delta wing and canard wing is achieved to perform the cruise model of the hybrid UAV.

In order to overcome the propre disadvantages of FW(Fixed-Wing) and VTOL(Vertical-Taking-Off-and-Landing) UAV (Unmanned Aerial Vehicle) and extend its application, the hybrid drone is invested more in recent years by researchers and several classifications are developed on the part of dual system. In this article, an innovative hybrid UAV is raised and studied by introducing the canard configuration that is coupled with conventional delta wing as well as winglet structure. Profited by Computational Fluid Dynamics (CFD) and Response Surface Method (RSM), a multilevel optimization approach is practically presented and concerned in terms of cruise flight mode: adopted by an experienced-based distribution strategy, the total lift object is respectively assigned into the delta wing (90–95%) and canard wing(5–10%) which is applied into a two-step optimization: the first optimization problem is solved only with the parameters concerned with delta wing afterwards the second optimization is successively concluded to develop the canard configuration considering the optimized delta wing conception. Above all, the optimal conceptual design of the delta and canard wing is realized by achieving the lift goal with less drag performance in cruise mode.