Search

Organizations widely use cloud computing to outsource their computing needs. One crucial issue of cloud computing is that services must be available to clients at all times. However, the cloud services may be temporarily unavailable due to maintenance of the cloud infrastructure, load balancing of services, defense against cyber attacks, power management, proactive fault tolerance, or resource usage. The unavailability of cloud services impacts negatively on the business model of cloud providers. One solution to tackle the service unavailability is Live Virtual Machine Migration (LVM), that is, moving virtual machines (VMs) from the source host machine to the destination host without disrupting the running application. Pre-copy memory migration is a common LVM approach used in most networked systems such as the cloud. The main difficulty with this approach is the high rate of frequently updating memory pages, referred to as "dirty pages. Transferring these updated or dirty pages during the pre-copy migration approach prolongs the total migration time. After a predefined iteration, the pre-copy approach enters the stop-and-copy phase and transfers the remaining memory pages. If the remaining pages are huge, the downtime or service unavailability will be very high -resulting in a negative impact on the availability of the running services. To minimize such service downtime, it is critical to find an optimal time to migrate a virtual machine in the pre-copy approach. To address the issue, this paper proposes a machine learning-based method to optimize pre-copy migration. It has mainly three stages (i) Feature selection (ii) Model generation and (iii) Application of the proposed model in pre-copy migration. The experiment results show that our proposed model outperforms other machine learning models in terms of prediction accuracy and it significantly reduces downtime or service unavailability during the migration process. [ABSTRACT FROM AUTHOR]

The number of cloud-, edge-, and Internet of Things (IoT)-based applications that produce sensitive and personal data has rapidly increased in recent years. The IoT is a new model that integrates physical objects and the Internet and has become one of the principal technological evolutions of computing. Cloud computing is a paradigm for centralized computing that gathers resources in one place and makes them available to consumers via the Internet. Despite the vast array of resources that cloud computing offers, real-time mobile applications might not find it acceptable because it is typically located far from users. However, in applications where low latency and high dependability are required, edge computing—which disperses resources to the network edge—is becoming more and more popular. Though it has less processing power than traditional cloud computing, edge computing offers resources in a decentralized way that can react to customers' needs more quickly. There has been a sharp increase in attackers stealing data from these applications since the data is so sensitive. Thus, a powerful Intrusion Detection System (IDS) that can identify intruders is required. IDS are essential for the cybersecurity of the IoT, cloud, and edge architectures. Investigators have mostly embraced the use of deep learning algorithms as a means of protecting the IoT environment. However, these techniques have some issues with computational complexity, long processing times, and poor precision. Feature selection approaches can be utilized to overcome these problems. Optimization methods, including bio-inspired algorithms, are applied as feature selection approaches to enhance the classification accuracy of IDS systems. Based on the cited sources, it appears that no study has looked into these difficulties in depth. This research thoroughly analyzes the current literature on intrusion detection and using nature-inspired algorithms to safeguard IoT and cloud/edge settings. This article examines pertinent analyses and surveys on the aforementioned subjects, dangers, and outlooks. It also examines many frequently used algorithms in the development of IDSs used in IoT security. The findings demonstrate their efficiency in addressing IoT and cloud/edge ecosystem security issues. Moreover, it has been shown that the methods put out in the literature might improve IDS security and dependability in terms of precision and execution speed. [ABSTRACT FROM AUTHOR]

Utilizing the cloud environment is one of the most preferable option in every information technology (IT) organization for running its business due to its flexible nature of services for its users. Cloud computing is vulnerable to various types of known and unknown attacks due to its distributed nature and open architecture. Hence, privacy and security is a primary concern of the cloud computing environment. A lot of machine learning approaches are utilized to improve the accuracy of Intrusion detection system (IDS) but dealing with redundant and non-relevant datasets with a large number of attributes (multi-dimensional) is still a problem. In this study, better and fast IDS has been proposed to detect anomaly for the cloud network environment which uses Improved squirrel search algorithm (ISSA) and Modified-deep belief network (MDBN) on the UNSW-NB15 dataset. ISSA extracts the relevant features from a set of features to deal with network traffic data of high dimension. It selects the relevant and optimal feature subset to be utilized for the training process and testing process. Meanwhile, a MDBN is introduced which uses the extracted features for anomaly detection for binary class classification and multi class classification. MDBN also helps to deal with the unbalanced nature of the dataset. The proposed IDS model has been evaluated based on the statistical measures namely accuracy, false alarm rate (FAR) and execution time, which are the popular parameters for evaluation of an IDS model. The proposed ISSA-MDBN model for IDS reduces the training from 103.51 to 0.108 s and testing time from 29.62 to 0.054 s on the UNSW-NB15 dataset. The proposed IDS model has been compared with other existing recent approaches and the proposed approach achieves the highest accuracy of 99.8% and lowest FAR of 0.02%. [ABSTRACT FROM AUTHOR]