Empirical analysis of machine learning-based energy efficient cloud load balancing architectures: a quantitative perspective

With the augmentation of affordable multimedia wireless gadgets, the ubiquitous availability of the internet access, and the rapid pace of mobile traffic motivate research towards fifth generation (5G) communications to realize energy-efficient cloud radio access networks (C-RAN) with guaranteed quality of experience. Exploiting green energy harvesting for powering the C-RAN substantially alleviates the energy procurement from the utility grid, carbon footprint, and operational expenses. In this paper, we propose a new dynamic point selection coordinated multipoint (DPS CoMP) based load balancing paradigm emphasizing achievable throughput and energy efficiency (EE) by reducing utility grid consumption from a network level perspective. This paper investigates the radio efficiency, EE, and average on-grid energy saving addressing the key challenges of tempo-spatial dynamics of traffic intensity and renewable energy (RE) generation under a wide range of network setup. Endeavoring load balancing technique strives a balance in network utilities such as green energy utilization and user association based on BS coordination technique in a cluster. Provision of cell sleep approach is contemplated for further energy saving by turning off lightly loaded base stations (BSs) during low traffic arrivals. The proposed CoMP based load balancing algorithm proficiently manages resource block allocation to the new users and elevated the energy efficiency over the conventional location and traffic centric mechanisms. Extensive system-level simulations manifest that the suggested framework enables an adjustable trade-off between radio efficiency and EE, and saves 22% on-grid power consumption and increases EE index by 32%. Afterward, an exhaustive comparison of the proposed method with the existing schemes is pledged for further validation highlighting sustainable 5G wireless systems.

Objectives: In this paper various task scheduling algorithm along with load distribution techniques investigated to ensure efficient mapping of tasks to resources and for coherent resource utilization in heterogeneous environment. Statistical Analysis: A detailed comparative approach is conferred for different job scheduling and load balancing algorithm available in the literature. Considered parameters for comparisons are scheduling approaches, techniques, findings, benefits, pros and cons. This study facilitates to identify the efficient scheduling and load balancing approach which optimizes resource use, minimizes completion time as well as task waiting time and balanced load among resources. Findings: Make span time and flow time are two important parameters to compute and minimize job execution time. This paper explores an intense survey of minimization of task execution time with minimum make span time and flow time. Apart from this a joint review on job scheduling, load balancing and minimizing job completion time using Expected Time to Compute (ETC) matrix is presented. At the end of this paper a research model is proposed which is a novel and hybrid approach with a combination of the efficiency of scheduling and load balancing. Application: The Load Balancing (LB) is an important feature of the grid environment. The efficient LB algorithm assures cost effective resource usage by the provisioning of resources to users on demand schedule in pay-as-you-say-manner.

Energy efficiency of wireless networks is becoming more and more important because of the steadily rising energy cost and environmental concerns. In a real network with unequal traffic load distributions among cells, by offloading some edge users in a heavily-loaded cell to neighboring lightly-loaded cell, load balance, aims to utilize the radio resource more efficiently and consequently can improve the end-user performance. In this paper an energy efficient load balancing algorithm is proposed. In the proposed algorithm, the target cell in load balance is selected based on the prediction of the possible performance. Furthermore, an energy efficiency scaling factor is introduced into the criteria when selecting the target cell in load balance to achieve the joint optimization of both load balance and energy saving. Simulation results show that the proposed algorithm can reduce the power consumption in the network and at the same time improve the cell throughput.

Cellular wirelesses heterogeneous networks (HetNets) have become an important part of the growth of the telecom industry. The rapid growth of cellular networks and widespread use of wireless services have led to a huge increase in energy use. This makes it hard to make cellular networks that use less energy and last longer. In HetNets, load balancing (LB) and energy balance have a big effect on cell range extension (CRE) and network lifetime extension (NLE). optimal energy efficient LB in wireless cellular HetNets is presented in this paper using a self-tuning algorithm. Several self-tuning methods for HetNets have been developed that are driven by the quality of experience (QoE). Unlike earlier heuristic-based approaches, the suggested technique uses a gradient ascending method to ensure that parameter modifications constantly increase overall system QoE, LB, throughput, and energy efficiency. The planned network's packet loss rate is extremely low. The suggested and contemporary models are tested in an experimental investigation, demonstrating the importance of the suggested model in comparison to other contemporary methodologies.

The future growth of the Internet of Services has fundamentally changed the emergence of cloud computing. Cloud data centres serve multiple tenant demands for cloud applications that discharge vast amounts of electricity, leading to high operating costs and environmental diffusion of carbon dioxide (CO2). To fix this is the need for preservation to enable potential use by building a new structure and measuring the effect in a cloud data centre. Consequently, the use of pruned electricity reduces the cost of processing power. In order to meet energy-efficient data centres in the cloud, adjusting to optimal load balancing processing a good way for energy savings. To minimise the large energy use of cloud data centres herds, this focuses on increasing efficacy by breaking the workload evenly. In this paper, we plan to provide a comprehensive comparative analysis in cloud computing of current load balancing algorithms.

Abstract: Enhancing route request broadcasting protocols constitutes a substantial part of recent research in mobile ad-hoc network (MANET) routing. We suggest a novel approach to modify route request broadcast based on node caching. Intuition behind node caching is that the nodes involved in recent data packet forwarding have more reliable information about its neighbors and have better locations (e.g., on the intersection of several data routes) than other nodes. We cache nodes which are recently involved in data packet forwarding, and use only them to forward route requests. Dropping route request forwarded from the other nodes considerably reduces routing overhead at the expense of possible destination missing. The suggested node caching techniques can be also viewed as a dynamic implementation of a connected dominating set (CDS). We overcome the known drawback of CDS – overuse of dominating (cached) nodes – by a new load-balancing scheme. Our contributions include: (i) a new energy-efficient node caching enhancement of route request broadcast for reactive ad hoc routing protocols; (ii) an extensive simulation study in NS2 of the novel node caching enhancement of AODV (AODV-NC) showing drastic reduction in overhead, significant improvement of the packet delivery ratio and the end-to-end delay and overhead; (iii) analysis of forwarding load distribution and energy consumption, and (iv) an extensive simulation study in NS2 of the novel AODV-NC based routing protocol with adaptive workload balancing (AODV-NC-WLB) sustaining considerable improvement in throughput, overhead, delivery ratio and delay over the standard AODV for stressed MANETs. Key Words: routing protocols, mobile ad hoc networks, Ad-hoc On-demand Distance Vector, routing load balancing, performance evaluation, node caching, energy efficiency.

ABSTRACTThe performance of 3GPP long term evolution (LTE) network can be significantly influenced by unbalanced user distribution. In heavily loaded cells, users cannot obtain satisfied services because of shortage of resources, whereas in lightly loaded cells, there are plenty of unoccupied resources. Load balancing (LB) is proposed to deal with this problem, through which the load distribution among nearby cells can be greatly balanced. However, because of the usage of full frequency reuse in LTE, the cell‐edge throughput is severely deteriorated by strong intercell interference. Meanwhile, LTE network usually serves users with heterogenous quality‐of‐service (QoS) requirements, which requires LB to be aware of it. Furthermore, call admission control (CAC) is indispensable in the process of LB, and it depends heavily on the long‐term average performance of the scheduling scheme. Thus, design of an efficient LB algorithm with the consideration of intercell interference, QoS, CAC and scheduling together becomes an important issue. To deal with the previous problem, a multi‐objective optimisation problem is formulated in this paper. Its complexity is analysed, and a practical algorithm, which includes hybrid scheduling, QoS‐guaranteed CAC, QoS‐aware intracell and intercell LB, is proposed. Extensive simulations are conducted. Results show that the proposed algorithm can approximate the optimal solution efficiently, thus leading to significantly better performance, that is, a lower new call blocking rate for users with QoS requirement, a larger total utility and a higher cell‐edge throughput for users without QoS requirement, although the cost is a bit degradation of the total throughput. Copyright © 2013 John Wiley & Sons, Ltd.

Cloud computing is an emerging computing paradigm in which shared resources are provided according to the customer request at specific time. Load balancing is the process of distributing workload among various nodes of the computing system. The load can be CPU load, memory capacity, or network load. An efficient load balancing avoids a situation where some of the nodes are heavily loaded while other nodes are idle or doing very little work. When Virtual Machine (VM) is overloaded with multiple tasks, these tasks are removed and migrated to the under loaded VMs of the same or different datacenter. This paper proposes a bee colony based algorithm for efficient load balancing, which is based on the foraging behavior of honey bees to balance load across VMs. In the proposed method, tasks removed from over loaded VMs are treated as honey bees and under loaded VMs are the food sources. The proposed method also considers the priorities of tasks in the waiting queues of VMs and tries to achieve minimum response time and reduced number of task migrations. The experimental result shows that there is significant improvement in the Quality of Service (QoS).

Resource Utilization Prediction with Multipath Traffic Routing for Congestion-aware VM Migration in Cloud Computing

The ever-increasing demand for cloud services results in large electricity costs to cloud providers and causes significant impact on the environment. This has pushed cloud providers to power their data centers with renewable energy sources more than ever. Among the different ways of adopting renewable energy sources, on-site power generation using wind and solar energy has gained considerable attention by large companies and proved its potential to reduce data centers' carbon footprint and energy costs. Efficient utilization of renewable energy sources is challenging due to their intermittency and unpredictability. Cloud providers with multiple Geo-distributed data centers in a region can exploit the temporal variations in on-site power and grid power price by routing the load to a suitable data center in order to reduce cost and increase renewable energy utilization. To achieve this goal, we propose a fuzzy logic-based load balancing algorithm that acts with no knowledge of future. We conduct extensive experiments using a case study based on real world traces obtained from National Renewable Energy Laboratory (NREL) and Energy Information Administration (EIA) in the US, and Google cluster-usage. Compared to other benchmark algorithms, our method is able to significantly reduce the cost without a priori knowledge of the future electricity price, renewable energy availability, and workloads.

Ensuring Quality of Service (QoS) and efficient load balancing in cloud communication networks is critical for optimizing resource allocation, minimizing latency, and enhancing service reliability. Traditional load balancing strategies often fail to scale and adapt to dynamic cloud environments, resulting in network congestion, resource underutilization, and increased operational costs. This study presents a novel Genetic Algorithm (GA)-based hyperheuristic optimization framework integrated with a mathematical model for QoS-aware load balancing, designed to address the challenges of scalability and efficiency. The model is referred to as GAHO\textsubscript{QoS}. We introduce valid inequalities to strengthen the optimization formulation, accelerating convergence and improving solution quality. Our GA-hyperheuristic framework dynamically selects and combines multiple low-level heuristics to optimize task allocation across cloud servers while adhering to QoS constraints such as latency, throughput, and energy efficiency. Experimental evaluations on a range of cloud communication scenarios demonstrate that GAHO\textsubscript{QoS} significantly reduces service latency, balances workload distribution, and optimizes resource utilization. Comparative analysis with existing metaheuristic methods, including $GA-PSO$ and $SA-GA$, confirms that the proposed framework outperforms traditional approaches in terms of computational efficiency, scalability, and QoS satisfaction. GAHO\textsubscript{QoS} provides an adaptable, computationally efficient solution for enhancing cloud network performance, contributing to the development of high-performance, energy-efficient, and robust cloud infrastructures.

Decades of techno-economic energy policymaking and research have meant evidence from the Social Sciences and Humanities (SSH)—including critical reflections on what changing a society’s relation to energy (efficiency) even means—have been underutilised. In particular, (i) the SSH have too often been sidelined and/or narrowly pigeonholed by policymakers, funders, and other decision-makers when driving research agendas, and (ii) the setting of SSH-focused research agendas has not historically embedded inclusive and deliberative processes. The aim of this paper is to address these gaps through the production of a research agenda outlining future SSH research priorities for energy efficiency. A Horizon Scanning exercise was run, which sought to identify 100 priority SSH questions for energy efficiency research. This exercise included 152 researchers with prior SSH expertise on energy efficiency, who together spanned 62 (sub-)disciplines of SSH, 23 countries, and a full range of career stages. The resultant questions were inductively clustered into seven themes as follows: (1) Citizenship, engagement and knowledge exchange in relation to energy efficiency; (2) Energy efficiency in relation to equity, justice, poverty and vulnerability; (3) Energy efficiency in relation to everyday life and practices of energy consumption and production; (4) Framing, defining and measuring energy efficiency; (5) Governance, policy and political issues around energy efficiency; (6) Roles of economic systems, supply chains and financial mechanisms in improving energy efficiency; and (7) The interactions, unintended consequences and rebound effects of energy efficiency interventions. Given the consistent centrality of energy efficiency in policy programmes, this paper highlights that well-developed SSH approaches are ready to be mobilised to contribute to the development, and/or to understand the implications, of energy efficiency measures and governance solutions. Implicitly, it also emphasises the heterogeneity of SSH policy evidence that can be produced. The agenda will be of use for both (1) those new to the energy-SSH field (including policyworkers), for learnings on the capabilities and capacities of energy-SSH, and (2) established energy-SSH researchers, for insights on the collectively held futures of energy-SSH research.

Domain specific allocation plays a major role in cloud computing. The recent trends in cloud computing attracts user to generate random requests to any processor which are not domain specific. Therefore load balancing is a significant aspect in distributing the workload evenly among all the nodes. Since the workload is not distributed uniformly, some processors are overloaded while others are under-loaded. This might result with the resources being overheated, sometimes might lead to carbon emission. Therefore in domain-specific storage text based and image based files are stored separately in two virtual machines for faster response. This paper presents the SJF scheduling algorithm for load balancing and AES algorithm for data encryption. The data is migrated into cloud called CloudMe for data recovery and data storage. This reduces the excessive work of the servers and provides efficient load balancing. The project handles both basic and premium users where premium user requests are prioritized and handled. The ‘n’ number of requests that is generated using bots by brute force attack can also be prevented. Thus the project aims to maintain a domain-specific storage to avoid overloading of virtual machines and maintains system stability and reduce carbon emission.

The most complicated process in multi‐cloud computing is resource allocation, as it needs to cope with a number of configurations and constraints of cloud providers and customers. At the time of resource allocation, the centralized cloud broker monitors the virtual machines (VM) status, scheduling process, and fitness. However, VM scheduling is found tedious and has received huge attention in business, academia, and research. This enhances the demand for both task scheduling and resource allocation in a multi‐cloud environment. To bridge the gap between the consumer requirement and server infrastructure, a joint optimization‐based resource allocation and task scheduling concept is analyzed in the proposed framework. The first phase introduces the task scheduling mechanism, which uses Hybrid Electro Search and Beetle Swarm Optimization to determine the optimal task for specific VMs. The optimal selection procedure is done by analyzing a multi‐cloud environment's makespan, energy, cost, and throughput parameters. In the second step, an Adaptive Game Theory‐based Seagull optimization approach performs several rounds of reassignment iteratively to minimize the variation in the expected completion time, consequently decreasing high energy consumption and load balancing. The experimental analysis for the proposed model is implemented using Python. The proposed methodology is shown to achieve cheaper costs, shorter waiting times, improved resource allocation, and efficient load balancing. Finally, a comparative analysis is performed with some hybrid optimization models, which illustrate the efficiency of the proposed hybrid optimization model.

The rapid development of cloud computing expands the scale of modern cloud data centers, resulting in growing challenges of energy consumption. Load balancing of virtual machines(VMs), for the purposes of improving the utilization of physical resources and reducing energy consumption, has become a research focus in recent years. However, the existing researches mostly focus on how to maximize resource utilization and reduce energy consumption. Security issues in the context of load balancing of VMs were rarely addressed. In this paper, we research the key procedures of load balancing, VM selection and VM placement, we find that the existing schemes introduced several security problems. In consideration of some conclusions of recent researches and inevitable live migrations during load balancing, common tenants have sufficient reasons to worry about their VMs' security when they are migrated to strange hosts and/or co-reside with the VMs owned by strange tenants. In short, VMs' mobility introduced by load balance expands the attack surface. In this work, we classify and analyze related security threats and create an information leakage model for load balancing. We present a new security policy, SeLance, to secure the load balancing via avoiding above threats as far as possible. We develop exact implementations in CloudSim and OpenStack. We show that SeLance can effectively alleviate the threats introduced by load balancing, the security score can improve 46.90%-81.15%, while keeping the load balancer's original function to a great extent (± 2.5%).