YSYassir Samadi

Yassir Samadi

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Yassir Samadi has a total of 9 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • DT-MG: many-to-one matching game for tasks scheduling towards resources optimization in cloud computing

    by Yassir Samadi, Mostapha Zbakh, Claude Tadonki

    The increasing demand of cloud computing motivates researchers to make cloud environments more efficient for its users and more profitable for the providers. More and more datacenters are being built to cater customers' needs. However, datacenters consume large amounts of energy, and this draws negative attention. Therefore, cloud providers are confronted with great pressures to reduce the energy consumed by datacenters. To address this issue, efficient algorithms to reduce energy consumption and to guarantee the quality of service are needed. In this paper, we propose a load balancing algorithm named DT-MG, which aims to reduce energy consumption and maximize the efficiency of the available resources. First, we used the Matching Game Theory model for assigning tasks to datacenters. We then study the optimal operation of the resources by migrating all the tasks of the physical machine under sub-regime to other physical machine, followed by their systematic switch to standby mode. Experimental results prove that the proposed approach reduces energy consumption and the number of task migration while maintaining the service level agreement in comparison with some existing techniques.


  • E-HEFT: Enhancement Heterogeneous Earliest Finish Time algorithm for Task Scheduling based on Load Balancing in Cloud Computing

    by Yassir Samadi, Mostapha Zbakh, Claude Tadonki

    Cloud computing is one of the most widely spreaded platforms for executing tasks through virtual machines as processing elements. However, there are various issues that need to be addressed in order to be efficiently utilized for workflow applications. One of the fundamental issues in cloud computing is related to task scheduling. Optimal scheduling of tasks in cloud computing is an NP-complete optimization problem, and many algorithms have been proposed to solve it. Furthermore, the existing algorithms fail to either meet the user's Quality of Service (QoS) requirements such as minimizing the makespan and satisfying budget constraints, or to incorporate some basic principles of cloud computing such as elasticity and heterogeneity of computing resources. Among these algorithms, the Heterogeneous Earliest Finish Time (HEFT) heuristic is known to give good results in short time for tasks scheduling in heterogeneous systems. Generally, the HEFT algorithm yields good tasks execution time, but its drawback is that there is no load balancing. In this paper, an enhancement of Heterogeneous Earliest Finish Time (E-HEFT) algorithm under a user-specified financial constraint is proposed to achieve a well balanced load across the virtual machines while trying to minimize the makespan of a given workflow application. To evaluate the performance of the enhancement algorithm, we compare our algorithm with some existing scheduling algorithms. Experimental results show that our algorithm outperforms other algorithms by reducing the makespan and improving load balance among virtual machines.


  • Performance comparison between Hadoop and Spark frameworks using HiBench benchmarks

    by Yassir Samadi, Mostapha Zbakh, Claude Tadonki

    Summary: Big Data has become one of the major areas of research for cloud service providers due to a large amount of data produced every day and the inefficiency of traditional algorithms and technologies to handle these large amounts of data. Big Data with its characteristics such as volume, variety, and veracity (3V) requires efficient technologies to process in real time. To solve this problem and to process and analyze this vast amount of data, there are many powerful tools like Hadoop and Spark, which are mainly used in the context of Big Data. They work following the principles of parallel computing. The challenge is to specify which Big Data's tool is better depending on the processing context. In this paper, we present and discuss a performance comparison between two popular Big Data frameworks deployed on virtual machines. Hadoop MapReduce and Apache Spark are used to efficiently process a vast amount of data in parallel and distributed mode on large clusters, and both of them suit for Big Data processing. We also present the execution results of Apache Hadoop in Amazon EC2, a major cloud computing environment. To compare the performance of these two frameworks, we use HiBench benchmark suite, which is an experimental approach for measuring the effectiveness of any computer system. The comparison is made based on three criteria: execution time, throughput, and speedup. We test Wordcount workload with different data sizes for more accurate results. Our experimental results show that the performance of these frameworks varies significantly based on the use case implementation. Furthermore, from our results we draw the conclusion that Spark is more efficient than Hadoop to deal with a large amount of data in major cases. However, Spark requires higher memory allocation, since it loads the data to be processed into memory and keeps them in caches for a while, just like standard databases. So the choice depends on performance level and memory constraints.


  • Performance evaluation of StarPU schedulers with preconditioned conjugate gradient solver on heterogeneous (multi-CPUs/multi-GPUs) architecture

    by Najlae Kasmi, Mostapha Zbakh, Yassir Samadi, Rachid Cherkaoui, Amine Haouari

    We consider the problem of scheduling sparse linear application on heterogeneous (CPUs/GPUs) platform. More specifically, we focus on the preconditioned conjugate gradient solver (PCG) since it exhibits the main features of such problems. Indeed, the relative performance of CPU and GPU highly depends on the sub-routine: GPUs are for instance much more efficient to process regular kernels such as matrix vector multiplications rather than more irregular kernels such as matrix factorization. In this context, one solution consists in relying on dynamic scheduling and resource allocation mechanisms such as the ones provided by StarPU. In this paper we evaluate the performance of dynamic schedulers proposed by StarPU, and we analyse the scalability of PCG algorithm. We show how effectively we can choose the best combination of resources in order to improve their performance.

  • TASMR: Towards advanced secure mapreduc framework across untrusted hybrid clouds

    by Amine Haouari, Mostapha Zbakh, Rachid Cherkaoui, Yassir Samadi, Najlae Kasmi

    MapReduce has become increasingly popular as a programming system for distributed processing large-scale data. It is largely used daily around the world as an efficient distributed computation tool for data processing in different research areas. To deploy it as a data processing service, privacy and security of MapReduce computations and data are crucial concerns when the it is executed over hybrid clouds. To this end we must provide necessary security mechanisms to protect MapReduce data processing jobs. Thus authentication of mappers-reducers, client-user, confidentiality and integrity of data-computations and freshness of the outputs are obligatory. Fulfilling these requirements safeguard MapReduce computations and data from various types of attacks. In this paper, we discuss security challenges and requirements to shield the computations and data. Hence we present a new architecture to secure MapReduce computation upon the aforementioned challenges in a hybrid clouds. Because often the distribution is carried on public cloud the user don't have a visibility upon the infrastructure nor what is happening during the processing. In the other hand the cloud provider should secure its infrastructure against malicious users. Thus we propose a novel architecture that takes the best from SecureMR, vTPM and Semrod. The proposed architecture not only overcomes the flaws of the state-of-the-art proposed solutions, but also provides remarkable security guarantees that guard against insider and outsider threats. We also compare and analyze the security overhead and features of our model with the state-of-the-art architectures already present.


  • Open Access

    Threshold-based load balancing algorithm for Big Data on a Cloud environment

    by Yassir Samadi, Mostapha Zbakh

    In this paper, we discuss a load balancing strategy in heterogeneous cloud environments. Load balancing of data processing takes a very important place in cloud computing for several years. Load balancing on cloud systems is critical problem that is difficult to cope with, especially on the emerging heterogeneous clusters. In this aspect, we propose a threshold-based load balancing algorithm that balances the load among datacenters in cloud environments as well as minimizing remote communication among datacenters. The proposed approach is divided into two phases. Firstly, we specify the load threshold of each datacenter based on its processing speed and storage capacity. Secondly, we maintain load balancing among datacenters based on this threshold while the proposed approach takes into consideration the heterogeneity of datacenters. The results show that our approach improve efficiently the load balancing among datacenters.


  • Comparative study between Hadoop and Spark based on Hibench benchmarks

    by Yassir Samadi, Mostapha Zbakh, Claude Tadonki

    Big Data is currently a hot topic for companies and scientists around the world, due to the emergence of new technologies, devices and communication means like social network sites, which led to a noticeable increase of the amount of data produced every year, even every day. In addition, traditional algorithms and technologies are inefficient to process, analyze and store this vast amount of data. So, to solve this problem, Big Data frameworks are needed. In this paper, we present and discuss a performance comparison between two popular Big Data frameworks. Hadoop and Spark, which are used to efficiently process vast amount of data in parallel and distributed mode on a large clusters. Hibench benchmark suite is used to compare the performance of these two frameworks based on the criteria as execution time, throughput and speedup. Our experimental results show that Spark is more efficient than Hadoop to deal with large amount of data. However, spark requires higher memory allocation, since it loads processes into memory and keeps them in caches for a while, just like standard databases. So the choice depends on performance level and memory constraints.