Samir Youcef

Bi-criteria Workflow Tasks Allocation and Scheduling in Cloud Computing Environments

Although there are few efficient algorithms in the literature for scientific workflow tasks allocation and scheduling for heterogeneous resources such as those proposed in grid computing context, they usually require a bounded number of computer resources that cannot be applied in Cloud computing environment. Indeed, unlike grid, elastic computing, such asAmazons EC2, allows users to allocate and release compute resources on-demand and pay only for what they use. Therefore, it is reasonable to assume that the number of resources is infinite. This feature of Clouds has been called âillusion of infiniteresourcesâ. However, despite the proven benefits of using Cloud to run scientific workflows, users lack guidance for choosing between multiple offering while taking into account several objectives which are often conflicting. On the other side, the workflow tasks allocation and scheduling have been shown to be NP-complete problems. Thus, it is convenient to use heuristic rather than deterministic algorithm. The objective of this paper is to design an allocation strategy for Cloud computing platform. More precisely, we propose three complementary bi-criteria approaches for scheduling workflows on distributed Cloud resources, taking into account the overall execution time and the cost incurred by using a set of resources.

Resources allocation and scheduling approaches for business process applications in Cloud contexts

Resources allocation and scheduling has been recognised as an important topic for business process execution. However, despite the proven benefits of using Cloud to run business process, users lack guidance for choosing between multiple offering while taking into account several objectives which are often conflicting. Moreover, when running business processes it is difficult to automate all tasks. In this paper, we propose three complementary approaches for Cloud computing platform. On the other side, elastic computing, such as Amazon EC2, allows users to allocate and release compute resources (virtual machines) on-demand and pay only for what they use. Therefore, it is reasonable to assume that the number of virtual machine is infinite while the number of human resources is finite. This feature of Clouds has been called “illusion of infinite resources”. In this paper, we design an allocation strategy for Cloud computing platform taking into account the above characteristics. More precisely, we propose three complementary bi-criterion approaches for scheduling business process on distributed Cloud resources.

Bi-criteria strategies for business processes scheduling in cloud environments with fairness metrics

The Cloud computing paradigm is adopted for its several advantages like reduction of cost incurred when using a set of resources. Despite the many proven benefits of using a Cloud infrastructure to run business processes, it is still faced with a major problem that can compromise its success: the lack of guidance for choosing between multiple offerings. To ensure this, we propose a set of algorithms for business process scheduling in Cloud computing environments. More precisely, we propose an extension of our previous approaches taking into account the fact that several instances of the same process can run simultaneously, and they may have to share the same resources. The proposed approaches take into account the Cloud elasticity feature on the one hand, and on the other hand, they consider the two most important quality of service criteria when running business process in Clouds environment, namely (i) the overall execution time and (ii) the cost incurred using a set of resources. In addition they allow to ensure fairness between the different concurrent business process instances.

Scheduling Strategies for Business Process Applications in Cloud Environments

The Cloud computing paradigm is adopted for its several advantages like reduction of cost incurred when using a set of resources. However, despite the many proven benefits of using a Cloud infrastructure to run business processes, it is still faced with a major problem that can compromise its success: the lack of guidance for choosing between multiple offerings. Moreover, when running business processes it is difficult to automate all tasks and several objectives often conflicting must be taken into account. For this, the authors propose a set of scheduling strategies for business processes in Cloud contexts. More precisely, the authors propose three bi-criteria complementary approaches for scheduling business processes on distributed Cloud resources while taking into account its elastic computing characteristic that allows users to allocate and release compute resources virtual machines on-demand and its business model based on pay as you go. Therefore, it is reasonable to assume that the number of virtual machine is infinite while the number of human resources is finite. Experiment results demonstrate that the proposed approaches present good performances.

Paving the Way towards Semi-automatic Design-Time Business Process Model Obfuscation

Business process (BP) stakeholders want to enjoy the benefits of the cloud, but they are also reluctant to expose their BP models which express the know-how of their companies. To prevent such a know-how exposure, this paper proposes a design-time approach for transforming a BP model into BP fragments so that these BP fragments externalized in a multi-cloud context do not allow a cloud resource provider to understand a critical fragment of the company. While existing contributions on this topic remain at the level of principles, we propose an algorithm supporting automatically such a BP model transformation.

A Multi-criteria Based Approach for Web Service Selection Using Quality of Service (QoS)

The development of the web technologies and the increasing of available services have introduced the issue of the selection of the most appropriate service among a set of candidate web services. First of all, the services offering a given functionality are discovered. Then, the service section process assists users in choosing the services that better meets their preferences. These preferences are generally, expressed as potentially objective functions often conflicting. Most of existing works trying to select the best web services are based either on a single evaluation criterion or, at best, on the use of an aggregation function like weighted sum of several quantitative evaluation criteria, or the use of the Pareto optimality notion. The work presented here addresses these shortcomings by introducing a new optimality notion based on two tests: (i) concordance and (ii) discordance tests. It presents an efficient algorithm to select only the best services using the introduced optimality notion. Moreover, the proposed algorithm exhibits encouraging results as supported by a series of experiments.

Bounding models families for performance evaluation in composite Web services

Abstract One challenge of composite Web service architectures is the guarantee of the Quality of Service (QoS). Performance evaluation of these architectures is essential but complex due to synchronizations inside the orchestration of services. We propose methods to automatically derive from the original model a family of bounding models for the composite Web response time. These models allow to find the appropriate trade-off between accuracy of the bounds and the computational complexity. The numerical results show the interest of our approach w.r.t. complexity and accuracy of the response time bounds.

Obfuscating a Business Process by Splitting Its Logic with Fake Fragments for Securing a Multi-cloud Deployment

Companies are ready to outsource their business process to the cloud to enjoy its benefits. However they must be sure that their business know-how is preserved. Although some solutions already exist and consist in splitting the model into a collaboration of BP fragments for a deployment in a muti-cloud context, none of them takes into account the possibility of collusion between the clouds executing the different fragments. To address this issue, we propose in this paper a solution consisting in adding fake BP fragments at specific locations in the process, so as to complicate its structure and hide the direct interaction between clouds executing sensitive fragments. Therefore, the discovery of the process by malicious cloud providers is delayed. The approach is validated against an introduced metric. It demonstrates that our approach is better in the worst case than previous approaches in best cases.

Business Process Design by Reusing Business Process Fragments from the Cloud

The constant development of technologies forces companies to be more innovative in order to stay competitive. In fact, designing a process from scratch is time consuming, error prone and costly. In this context, companies are heading to reuse process fragments when designing a new process to ensure a high degree of efficiency, with respect to delivery deadlines. However, reusing these fragments may disclose sensitive business activities, especially if these latter are deployed in an untrusted environment. In addition, companies are concerned about their users privacy. To address these issues, we investigate how to build a new business process by reusing the safest existing fragments coming from various cloud servers, i.e. The ones that comply at best with companys preferences and policies, and offer an appropriate level of safety.

Business Process Scheduling Strategies in Cloud Environments with Fairness Metrics

Matching and scheduling problem proved to be a critical problematic in different domains including Cloud computing. Therefore, to ensure the commercial success of the Cloud computing paradigm, it is necessary to develop methods that allow users to optimize the use of resources. Even though there are several algorithms for scheduling applications in heterogeneous environment such as grid computing, they cannot benefit from the recent advent of Cloud computing. Indeed, these algorithms assume that the number of resources available to users is bounded, this is against the illusion of infinite resources of Cloud computing. Also, only the execution time (makespan) is taken into account. However, Cloud computing business model is based on pay as you go. Accordingly, execution cost begot using a set of resources should be considered. To overcome the limitations of existing works, this paper propose new strategies for matching and scheduling business process instances in the Cloud context. The proposed strategies aim at scheduling business process instances while minimizing two conflicting criteria on the one hand, and ensuring fairness between the considered instances on the other hand. A serie of experiments demonstrate that they present good performances.