Smruti Rekha

The rapid growth of geographically distributed cloud data centers has intensified the demand for secure, privacy-preserving, and resource-efficient collaboration among mutually untrusted cloud environments. This paper presents BlockFed, a blockchain-empowered federated learning framework designed to enable cyber-threat-resilient collaboration across geographically distributed data centers. In BlockFed, each data center independently trains local models using private workload data and shares only the computed gradients rather than raw data, with a centralized aggregation server through a secure blockchain layer. A Proof-of-Work consensus mechanism is employed to validate gradient transactions and maintain an immutable, tamper-resistant ledger, ensuring trust and integrity among participating entities. The centralized server aggregates blockchain-verified updates to construct a global model, which is iteratively redistributed to support collaborative learning. This integrated learning process enhances task-level resource demand prediction while simultaneously improving system security and operational efficiency across all participating data centers. Extensive simulations on the Google Cluster Dataset show that BlockFed outperforms state-of-the-art baselines, including ISTM, ETP-WE, First-Fit, Best-Fit, and Random-Fit, in resource utilization, power consumption, and active server reduction. BlockFed achieves up to 7-81% higher resource utilization, 41-58% lower power consumption, and 31-65% fewer active servers, while attaining a cyber-threat estimation accuracy of 93.19%.

Secure resource management (SRM) within a cloud computing environment is a critical yet infrequently studied research topic. This article provides a comprehensive survey and comparative performance evaluation of potential cyber threat countermeasure strategies that address security challenges during cloud workload execution and resource management. Cybersecurity is explored specifically in the context of cloud resource management, with an emphasis on identifying the associated challenges. The cyber threat countermeasure methods are categorized into three classes: defensive strategies, mitigating strategies, and hybrid strategies. The existing countermeasure strategies belonging to each class are thoroughly discussed and compared. In addition to conceptual and theoretical analysis, the leading countermeasure strategies within these categories are implemented on a common platform and examined using two real-world virtual machine (VM) data traces. Based on this comprehensive study and performance evaluation, this article discusses the tradeoffs among these countermeasure strategies and their utility, providing imperative concluding remarks on the holistic study of cloud cyber threat countermeasures and SRM. Furthermore, the study suggests future methodologies that could effectively address the emerging challenges of secure cloud resource management.

Modern large-scale computing systems improve job completion and energy efficiency by executing tasks in parallel. However, the Long Tail problem where a few slow-performing tasks, or stragglers, delay overall execution remains a key challenge in cloud computing. This paper introduces an Energy-aware Straggler Task Management Framework for sustainable cloud environments. The framework integrates predictive analytics with an evolutionary Neural Network to identify stragglers based on historical resource usage. A self-adaptive differential evolution algorithm is employed to optimize the neural network weights, enhancing prediction accuracy. Tasks are classified as stragglers or non-stragglers, with stragglers further divided into Resource Hunter and Long-Tail categories. A resource management policy dynamically assigns resources based on task type to enhance parallelism and sustainability. Extensive simulations were performed using the Google Cluster Dataset to analyze the effectiveness of the EaSTM framework. The obtained results highlight the framework's potential for improving efficiency and sustainability in cloud environments.

Large-scale computing systems in the modern era distribute tasks into smaller units that can be executed simultaneously to speed up job completion and decrease energy usage. However, cloud computing systems encounter a significant challenge called the Long Tail problem, where a small subset of slow-performing tasks hinders the overall progress of parallel job execution. This behavior leads to longer service response times and reduced system efficiency. This paper introduces a novel approach called Stochastic Gradient Descent with Momentum-driven Neural Network to analyze and classify heterogeneous tasks as either stragglers or non-stragglers. The straggler tasks are further categorized into Resource Hunter and Long-Tail stragglers based on their specific resource requirements. A traffic management policy is implemented to schedule and assign resources among user job requests, considering the task category, to achieve parallelism and improve sustainability within the cloud infrastructure. Extensive simulations are conducted using the Google Cluster Dataset (GCD) to assess the effectiveness of the proposed framework. The results obtained from these simulations are then compared to state-of-the-art techniques. The experimental findings demonstrate significant reductions in power consumption, carbon emissions, active servers, conflicting servers, and VM migration up to 55.16%, 49.76%, 35%, 25.7%, and 87.29%, respectively. Moreover, there has been an enhancement in resource utilization by up to 78.31%, accompanied by a decrease in execution time of up to 67.74%.

The multi-tenant resource-sharing environment of cloud computing among multiple users and vulnerabilities of virtual machines (VM)s and physical machines prompt unauthorized access to users’ confidential data and excess power consumption and resource wastage. This paper presents a comparative survey study of securing cloud resource strategies to address this critical issue. Specifically, this study focuses on Defensive Strategies associated with cloud resource allocation with their pandect comparative summary.

Cloud computing (CC) has become the best way to manage a large-scale computing paradigm. Resource management (RM) in the cloud environment is critical due to the heterogeneity of computing resource types and their inter-dependencies. Recently, significant research has been done on RM techniques focusing on efficiently sharing resources among multiple users. Machine learning techniques manage various resource management functions, such as VM consolidation, resource optimization, energy optimization, workload estimation, and task scheduling. This article contributes a comprehensive and concise review of RM approaches within cloud environments. Moreover, it suggests design goals and research challenges that must be considered while developing novel RM strategies.

This letter proposes a novel Cloud Radio Access Network (C-RAN) traffic analysis and management model that estimates probable RAN traffic congestion and mitigate its effect by adopting a suitable handling mechanism. A computation approach is introduced to classify heterogeneous RAN traffic into distinct traffic states based on bandwidth consumption and execution time of various job requests. Further, a cloud-based traffic management is employed to schedule and allocate resources among user job requests according to the associated traffic states to minimize latency and maximize bandwidth utilization. The experimental evaluation and comparison of the proposed model with state-of-the-art methods reveal that it is effective in minimizing the worse effect of traffic congestion and improves bandwidth utilization and reduces job execution latency up to 17.07% and 18%, respectively.

This letter proposes a novel Cloud Radio Access Network (C-RAN) traffic analysis and management model that estimates probable RAN traffic congestion and mitigate its effect by adopting a suitable handling mechanism. A computation approach is introduced to classify heterogeneous RAN traffic into distinct traffic states based on bandwidth consumption and execution time of various job requests. Further, a cloud-based traffic management is employed to schedule and allocate resources among user job requests according to the associated traffic states to minimize latency and maximize bandwidth utilization. The experimental evaluation and comparison of the proposed model with state-of-the-art methods reveal that it is effective in minimizing the worse effect of traffic congestion and improves bandwidth utilization and reduces job execution latency up to 17.07% and 18%, respectively.

Security is always a prime concern of every cloud user. Many pioneer surveys have been done concerning security during data communication among owners, users, and stakeholders. However, there is a lack of literature surveys that address the concern of security during data processing or the workload execution within cloud data centers. In this context, we present a concise review of the existing methods that have addressed the security of workload or user data during load scheduling and execution within cloud environments. These methods have applied load-balancing strategies that helped to minimize the security attacks via co-residency or covert channels. The pandect comparative summary of these methods follows the corresponding review. The paper is concluded with limitations and future direction concerning secure load management.

Virtual machine placement (VMP) is the process of selecting the most appropriate physical machine (PM) to place users' requested virtual machine (VM) in large cloud data centers. Several methods have been framed to deal with this problem. However, the current solutions only consider limited resource types, resulting in an unbalanced load that activates unnecessary PMs inside the data center. In this article, we suggest a flower pollination-based nondominated sorting optimization (FP-NSO) algorithm that maximizes resource usage and minimizes energy consumption and carbon emission of the data center. Multiple resource-constraint metrics are associated with our algorithm that assists in finding the most suitable PMs for deploying VMs in a cloud environment. The VMP is carried out by employing the combined concept of flower pollination optimization and nondominated sorting technique-based genetic algorithm (NSGA-II). The algorithm is evaluated using the Google cluster dataset. The performance metrics like resource utilization, power consumption, and carbon emission values are computed for static and dynamic scenarios. The obtained results are compared with existing approaches. There is a significant reduction in power consumption, carbon emission, and execution time up to 16.69%, 48.60%, and 75.87%, respectively, and an improvement in resource utilization is up to 78.18%.