Search

Highlights • Heterogeneous cloud resources and the input tasks of a cloud system are modeled, and an energy-aware resource allocation platform is introduced to optimize the energy consumption of cloud data centers. • Propose a task-based VM-placement algorithm (ETVMC) to reduce the energy consumption, minimize the makespan of the system, and reduce the task rejection rate. • An experimental evaluation to validate the proposed solution utilizing the CloudSim as simulation framework. Abstract Employing cloud computing to acquire the benefit of cloud by optimizing various parameters that meet changing demands is a challenging task. The optimal mapping of tasks to virtual machines (VMs) and VMs to physical machines (PMs) (known as VM placement) problem are necessary for advancing energy consumption and resource utilization. High heterogeneity of tasks as well as resources, great dynamism and virtualization make the consolidation issue more complicated in the cloud computing system. In this paper, a complete mapping (i.e., task VM and VM to PM) algorithm is proposed. The tasks are classified according to their resource requirement and then searching for the appropriate VM and again searching for the appropriate PM where the selected VM can be deployed. The proposed algorithm reduces the energy consumption by depreciating the number of active PMs, while also minimizes the makespan and task rejection rate. We have evaluated our proposed approach in CloudSim simulator, and the results demonstrate the effectiveness of the proposed algorithm over some existing standard algorithms. [ABSTRACT FROM AUTHOR]

Data processing in real-time brings better business modeling and an intuitive plan of action. Internet of things (IoT), being a source of sensitive data collected and communicated through either public or private networks, requires better security from end to end to uphold integrity, quality, and acceptability of data. Designing an adaptive solution plays a vital role where IoT is deployed for the sensing-as-a-services in the critical infrastructure and near real-time decision making by deploying data analysis in the edge datacenters. Again, securing the system with user's demand and device specifications is a challenging and open research problem. This paper proposed a decision tree based user-centric security approach named D e c i s i o n T S e c that provides a secure channel for communication in IoT networks, combining edge datacenters in the network edges. Further, the proposed D e c i s i o n T S e c is validated by experimenting with the real-time testbed for the system performance along with the theoretical security validation. [ABSTRACT FROM AUTHOR]

• An operation on graphs is defined that obeys the properties of an algebraic group. • The proposed mechanism represents plaintext, ciphertext, and key as a graph. • It provides an unambiguous mapping process from plaintext to graph and vice versa. • Unlike other symmetric cryptosystems, the encryption and decryption keys are not the same. • The decryption key graph is the inverse graph of the encryption key graph. Data confidentiality and integrity are essential security goals in the data communication in the smart city or similar applications. Despite significant changes in modern data communication systems and devices, the cryptosystems are still the primary option to achieve data security. Several cryptosystems have been developed, implemented and standardized to date. However, there is always a continuous demand for a new cryptosystem for providing security during data communication. In this paper, a novel symmetric cryptosystem is presented for achieving data confidentiality. The proposed cryptosystem is a graph-based system, where an algebraic structure of directed graphs along with a novel operation is defined to provide the base of the cryptosystem. Here, a directed graph represents a plaintext, a ciphertext, and a secret key as well. The defined operation is used for encrypting and decrypting plaintext and ciphertext graphs. The proposed cryptosystem provides a simplified transformation process and representation of plaintext to plaintext graph. The correctness of the proposed cryptosystem is validated, and also shown that the conventional brute-force approach to obtain the key from either plaintext or ciphertext is exponential and this computational time is significantly higher than AES. This implies that the proposed cryptosystem is highly secured against this attack. The proposed cryptosystem is implemented and compared with AES and the results infer that it incurs a marginally higher execution time than the AES for larger plaintext due to the involvement of matrix operations. [ABSTRACT FROM AUTHOR]