A Krill Herd Behaviour Inspired Load Balancing of Tasks in Cloud Computing

Abstract

A developing trend in the IT environment is mobile cloud computing (MCC) with colossal infrastructural and resource requirements. In the cloud computing environment, load balancing -a way of distributing workloads across numerous computing resources, is a vital aspect. A proficient load balancing guarantees an effective resource usage through the supply of network resources based on the user demands. It can also organize the network clients using the fitting planning criteria. This paper sets forth an advanced load balancing and energy/cost aware technique for a demand-based network resource allocation in cloud computing. The load balancing process in the proposed strategy utilizes a Krill load balancer (Krill LB) which is expected to achieve a well-balanced load over virtual machines. The aim of using the Krill LB as the load balancer is to increase the throughput of the network as much as possible. The speed, task cost, and weight of the tasks were first determined, after which, the Krill herd optimization algorithm was for the load balancing based on the measured parameters. Furthermore, a modified dynamic energy-aware cloudlet-based mobile cloud computing model (MDECM) was introduced for energy cost awareness in load balancing based on the service rate and energy of the mobile users. The proposed work was aimed at optimizing resource allocation in MCC in an energy-efficient manner. The performance of the suggested Krill-LB was benchmarked against that of Honey Bee Behavior Load Balancing (HBB-LB), Kill Herd, and Round Robin algorithms. or applications to servers in cloud domain in order to execute them, and after that, recover the outcome of the execution from these servers [8]. It utilizes communication technology to share information and assets and incorporates locationaware technologies, mobile access to IT, and energy sparing technology specifically designed for mobile devices [15]. As of late, mobile applications have been noticeably copious with different classes, such as entertainment, health, games, business, social networking, travel, andIn the cloud, the load assigned to every node in the network is similarly distributed with an even quantity of resources over time. This enhances the scheme performance by moving the workloads among various nodes [4] [26]. The primary goal is to expedite the implementation of applications on resources whose workload changes at the runtime in an unpredictable way. These are generally discussed in homogeneous conditions such as grids. Fundamentally, there are two ways of load balancing procedures: (i) Static and (ii) dynamic [3]. The load balancing plan and a migration policy are aimed at virtual machine (VM) clustering to brilliantly choose a VM from an over-burdened Studies in Informatics and Control, 26(4) 413-424,

Efficient method for breast cancer classification based on ensemble hoffeding tree and naïve Bayes

“…This is a recent development [14][15][16][17][18] in data classification which performs prediction by selecting the majority class at each leaf. The incorporation of Naive Bayes models at the tree leaves can improve the predictive accuracy of HT.…”

Analysis of LTE physical channels overhead

“…Depending on the location of RBs, it carry different types of traffic in the frequency time matrix. Cell reference signals (CRS) is transmitted in a predetermined RBs and used in DL channel estimation by different US's [14,23,24]. CRSs are included for each antenna.…”

A beamforming study of the linear antenna array using grey wolf optimization algorithm

“…c) The factor " " is linearly reduced from 2 to 0, in order to confirm the searching and raiding of the prey, respectively. To avoid recession in local solutions, the candidate solutions will tend to converge towards the prey when | | < 1 and diverge from the prey when | | > 1 and [5,11]. d) Eventually, the GWO algorithm is stopped by the contentment of an end criterion.…”

“…The main purpose of beamforming is used to steer multiple beams towards wanted users while cancelling the interferers at the same time [10][11][12]. It can do that through adjustment of the beamformers weight vectors, where the quality of the communication channel can maximize through the process of altering the complex weight [13][14][15][16][17].…”