ANN-based secure task scheduling in computational clouds / Jacek Tchórzewski, Ana Respício, Joanna Kołodziej // W: 32nd European Conference on Modelling and Simulation ECMS 2018, May 22nd – May 25th, ... więcej

ANN-based secure task scheduling in computational clouds / Jacek Tchórzewski, Ana Respício, Joanna Kołodziej // W: 32nd European Conference on Modelling and Simulation ECMS 2018, May 22nd – May 25th, 2018, Wilhelmshaven, Germany : proceedings [Dokument elektroniczny] / ed. by Lars Nolle [et al.] ; org. by ECMS - European Council for Modelling and Simulation. – Sbr.-Dudweiler : European Council for Modelling and Simulation, 2018. – (European Conference for Modelling and Simulation, ISSN 2522-2422). – S. 468-474 : rys., tab. – Tryb dostępu: http://www.scs-europe.net/dlib/2018/2018-0468.htm. – Bibliogr. 19 poz., Abstr. – doi: 10.7148/2018-0468. – ISBN 978-0-9932440-6-3 (print). – ISBN 978-0-9932440-7-0 (CD)

Assuring the security of services in Computational Clouds (CC) is one of the most critical factors in cloud computing. However, it can complicate an already complex environment due to the complexity of ... więcej

Assuring the security of services in Computational Clouds (CC) is one of the most critical factors in cloud computing. However, it can complicate an already complex environment due to the complexity of the system architecture, the huge number of services, and the required storage management. In real systems, some security parameters of CC are manually set, which can be very time-consuming and requires security expertise. This paper proposes an intelligent system to support decisions regarding security and tasks scheduling in cloud services, which aims at automating these processes. This system comprises two different kinds of Artificial Neural Networks (ANN) and an evolutionary algorithm and has as main goal sorting tasks incoming into CC according to their security demands. Trust levels of virtual machines (VMs) in the environment are automatically set to meet the tasks security demands. Tasks are then scheduled on VMs optimizing the makespan and ensuring that their security requirements are fulfilled. The paper also describes tests assessing the best configurations for the system components, using randomly generated batches of tasks. Results are presented and discussed. The proposed system may be used by CC service providers and CC consumers using Infrastructure as a Service (IaaS) and Platform as a Service (PaaS) Cloud Computing models.