Load Balancing Issues Research Articles

Centralized SON function for operator optimal strategies in heterogeneous networks

Recently emerging fourth generation networks, notably LTE-Advanced (LTE-A), are expected to meet the requirements of higher bit rates with excellent quality of service (QoS). The expansion and heterogeneity of these networks have made their operational cost higher, resulting in harder tasks for operators to remain competitive. Therefore, automatic engineering has recently been considered as a promising method in remote network management and human intervention minimization. Self-organizing networks (SONs) offer the possibility of remotely controlling the network in both pre-operational and operational states through self-configuration, self-optimization, and self-healing. In this paper, we propose a global handover self-optimization (HSOP) algorithm that jointly treats load balancing, energy saving, and ICI issues and can be deployed in a dynamic, automatic, and centralized way. Our proposed HSOP algorithm allows the mobile operator to choose the best HSOP strategy according to the network characteristics (place, time, user equipment type, and priority, etc.). Therefore, a traffic model must be available for the operator to indicate the input parameters of our global SON algorithm.

Improvement of Adaptive Load Balanced Gateway Discovery Protocol in Hybrid Integrated Internet-MANET

<p class="0abstract">Internetworking of different types of networks is envisaged as one of the primary objectives of the future 5G networks. Integrated Internet-MANET is a heterogeneous networking architecture which is the result of interconnecting wired Internet and wireless MANET. Multiprotocol gateways are used to achieve this interconnection. There are two types of Integrated Internet-MANET architectures, two-tier and three-tier. A combination of two-tier and three tier architectures also exists, called the Hybrid Framework or Hybrid Integrated Internet-MANET. Some of the most important issues common to all Integrated Internet-MANET architecture are: efficient gateway discovery, mobile node registration and gateway load balancing. Adaptive WLB-AODV is an existing protocol which addresses the issues of Gateway load balancing and efficient Gateway discovery. In this paper, an improvement is proposed to Adaptive WLB-AODV, called Adaptive Modified-WLV-AODV by taking into account route latency. The proposed protocol has been implemented in Hybrid Integrated Internet-MANET and has been simulated using network simulation tool ns-2. Based on the simulation results, it is observed that the proposed protocol delivers better performance than the existing protocol in terms of performance metrics end-to-end delay and packet loss ratio. The performance of the proposed protocol is further optimized using a genetic algorithm.</p>

Load balancing, cost and response time minimisation issues in agent-based multi cloud service composition

Load balancing, cost and response time minimisation issues in agent-based multi cloud service composition

Load balancing, Cost and Response Time Minimization Issues in Agent Based Multi Cloud Service Composition

In multi cloud environments, we need to find services from multiple clouds, if a single cloud cannot give us all the required component services. In such a situation, however, it is challenging to find an appropriate composition sequence by taking into consideration load balancing between the different clouds and in the same time by minimising the composite service cost and response time. In this paper, an agent-based cloud service composition architecture is presented. It includes two main agents CSCA and CMA which use MCDA methods to support cloud services selection. These agents use load balancing mechanisms and the CNP in order to evolve and adapt cloud service composition. We have performed a comprehensive analytical and experimental study to evaluate the effectiveness of our approach. The experimental results, based on CloudSim simulator, show that the proposed architecture can effectively achieve good performance (load balancing), improve the response time and minimise cost.

An Efficient Parallel Method for Mining Frequent Closed Sequential Patterns

Mining frequent closed sequential pattern (FCSPs) has attracted a great deal of research attention, because it is an important task in sequences mining. In recently, many studies have focused on mining frequent closed sequential patterns because, such patterns have proved to be more efficient and compact than frequent sequential patterns. Information can be fully extracted from frequent closed sequential patterns. In this paper, we propose an efficient parallel approach called parallel dynamic bit vector frequent closed sequential patterns (pDBV-FCSP) using multi-core processor architecture for mining FCSPs from large databases. The pDBV-FCSP divides the search space to reduce the required storage space and performs closure checking of prefix sequences early to reduce execution time for mining frequent closed sequential patterns. This approach overcomes the problems of parallel mining such as overhead of communication, synchronization, and data replication. It also solves the load balance issues of the workload between the processors with a dynamic mechanism that re-distributes the work, when some processes are out of work to minimize the idle CPU time.

ARAM: A New Auction-based Resource Allocation Model in Cloud Computing

Cloud computing is an on-demand network access model to a shared pool of configurable computing resources. One of the main challenges in cloud computing is the efficiency of the pricing and resource allocation models adopted by cloud providers. In this paper, we proposed a new resource allocation model called ARAM, which is based on reverse auctions and provides dynamic pricing policies. In this model, independent reverse auctions are hold for each cloud service request. ARAM has been compared with other previously proposed approaches, considering several issues such as the types of the services it supports, the necessity for configuring service instances, and also those related to load balancing. The main advantage of ARAM, compared to previous forward auction based models, is that it eliminates the necessity for configuring service instances. Therefore, the utilities of both consumers and providers are enhanced, in terms of cost and time. Also by adopting a dynamic pricing approach, it moderates the shortcomings of fixed pricing strategies in cloud computing, such as resource wastage and lack of fairness. Furthermore, contrary to most of the previous approaches, ARAM supports dynamic pricing for all types of services and considers load balancing issues in allocating resources. Today, most of the providers in the industry use fixed pricing models ignoring the market condition. Therefore, it seems necessary to use efficient dynamic pricing and resource allocation models, such as ARAM, in cloud computing.

DLRankSVM: an efficient distributed algorithm for linear RankSVM

Linear RankSVM is one of the widely used methods for learning to rank. The existing methods, such as Trust-Region Newton (TRON) method along with Order-Statistic Tree (OST), can be applied to train the linear RankSVM effectively. However, extremely lengthy training time is unacceptable when using any existing method to handle the large-scale linear RankSVM. To solve this problem, we thus focus on designing an efficient distributed method (named DLRankSVM) to train the huge-scale linear RankSVM on distributed systems. First, to efficiently reduce the communication overheads, we divide the training problem into subproblems in terms of different queries. Second, we propose an efficient heuristic algorithm to address the load balancing issue (which is a NP-complete problem). Third, using OST, we propose an efficient parallel algorithm (named PAV) to compute auxiliary variables at each computational node of the distributed system. Finally, based on PAV and the proposed heuristic algorithm, we develop DLRankSVM under the framework of TRON. The extensive empirical evaluations show that DLRankSVM not only can obtain impressive speedups on both multi-core and distributed systems, but also can perform well in prediction compared with the other state-of-the-art methods. To the best of our knowledge, this is the first research work proposed to train the huge-scale linear RankSVM in a distributed fashion.

Dynamic Threshold Based Load Balancing Algorithms

In this scenario, dynamic and decentralized Load Balancing (LB) considers all the factors pertaining to the characteristics of the Grid computing environment. Dynamic load-balancing algorithms attempt to use the run-time state information to make more informative decisions in sharing the system load and in decentralization, algorithm is executed by all nodes in the system and the responsibility of LB is shared among all the nodes in the same pool. For this purpose, in this work, an extensive survey of the existing LB has been done. A detailed classification and gap analysis of the existing techniques is presented based on different parameters. The issue of LB in a Grid has been addressed while maintaining the resource utilization and response time for dynamic and decentralized Grid environment. Here, a hierarchical LB technique has been analyzed based on variable threshold value. The load is divided into different categories, like, lightly loaded, under-lightly loaded, overloaded, and normally loaded. A threshold value, which can be found out using load deviation, is responsible for transferring the task and flow of workload information. In order to improve response time and to increase throughput of the Grid, a random policy has been introduced to reduce the resource allocation capacity etc. Poisson process has been used for random job arrival and then load calculation has been done for assigning job to the appropriate Processing Entity for balancing the load in the pool. After balancing the load, it comes into the normally loaded pool, and then Job Migration process is executed. The performance of the proposed model, algorithms and techniques has been examined over the GridSim simulator using various parameters, such as response time, resource allocation efficiency, etc. Experimental results prove the superiority of the proposed techniques over the existing techniques.

OLB: A Nature Inspired Approach for Load Balancing in Cloud Computing

Abstract Load balancing is treated as one of the important mechanisms for efficient resource allocation in cloud computing. In future there will appear a necessity of fully autonomic distributed systems to address the load balancing issues. With reference to this, we proposed a load balancing mechanism called Osmosis Load Balancing (OLB). OLB works on the principle of osmosis to reschedule the tasks in virtual machines. The solution is based on the Distributed Hash Table (DHT) with a chord overlay mechanism. The Chord overlay is used for managing bio inspired agents and status of the cloud. By simulation analysis, the proposed algorithm has shown better performance in different scenarios, both in heterogeneous and homogeneous clouds.

Optimization and large scale computation of an entropy-based moment closure

We present computational advances and results in the implementation of an entropy-based moment closure, MN, in the context of linear kinetic equations, with an emphasis on heterogeneous and large-scale computing platforms. Entropy-based closures are known in several cases to yield more accurate results than closures based on standard spectral approximations, such as PN, but the computational cost is generally much higher and often prohibitive. Several optimizations are introduced to improve the performance of entropy-based algorithms over previous implementations. These optimizations include the use of GPU acceleration and the exploitation of the mathematical properties of spherical harmonics, which are used as test functions in the moment formulation. To test the emerging high-performance computing paradigm of communication bound simulations, we present timing results at the largest computational scales currently available. These results show, in particular, load balancing issues in scaling the MN algorithm that do not appear for the PN algorithm. We also observe that in weak scaling tests, the ratio in time to solution of MN to PN decreases.

Toward cluster-based weighted compressive data aggregation in wireless sensor networks

Conventional Compressive Sampling (CS)-based data aggregation methods require a large number of sensor nodes for each CS measurement leading to an inefficient energy consumption in Wireless Sensor Networks (WSNs). To solve this problem, we propose a new scheme in the network layer, called “Weighted Compressive Data Aggregation (WCDA)”, which benefits from the advantage of the sparse random measurement matrix to reduce the energy consumption. The novelty of the WCDA algorithm lies in the power control ability in sensor nodes to form energy efficient routing trees with focus on the load-balancing issue. In the second part, we present another new data aggregation method namely “Cluster-based Weighted Compressive Data Aggregation (CWCDA)” to make a significant reduction in the energy consumption in our WSN model. The main idea behind this algorithm is to apply the WCDA algorithm to each cluster in order to reduce significantly the number of involved sensor nodes during each CS measurement. In this case, candidate nodes related to each collector node are selected among the nodes inside one cluster. This yields in the formation of collection trees with a smaller structure than that of the WCDA algorithm. The effectiveness of these new algorithms is evaluated from the energy consumption, load balancing and lifetime perspectives of the network. A comprehensive numerical evaluation is performed which shows that the performance of the proposed WCDA and CWCDA algorithms is significantly better than some existing data aggregation methods such as plain-CS, hybrid-CS and the Minimum Spanning Tree Projection (MSTP) schemes.

QOS Differentiation of Various Cloud Computing Load Balancing Techniques

With an increase in the demands the Cloud computing has become one of the ongoing scalable approach to fulfill the cloud based neccesties. The biggest advantage of the cloud computing is the ability to overcome the infrastructural challenges those are earlier faced by other technologies. Since the technology is new, therefore the development of the whole structure is not so efficient. It does have a lot of issues on which various scientists and others are working on. Scheduling, load balancing, fault tolerance, etc. are various challenges faced by cloud computing. For this purpose various techniques and algorithms have been proposed. In this paper discussion will be done on the issue of load balancing in cloud computing and different types of load balancing techniques have been reviewed and a comparative analysis have been made among all the existing techniques. General Terms Cloud computing,Load balancing,Algorithms.

LOAD BALANCING ISSUES AND ITS SOLUTION IN CLOUD COMPUTING: A REVIEW

Cloud computing is an emerging paradigm in the computer industry where the computing is moved to a cloud of computers. Cloud computing is a way to increase the capacity or add capabilities dynamically without investing in new infrastructure, training new personnel, or licensing new software. This paper is focused on the load balancing issues of cloud computing and techniques to overcome the waiting time and turnaround time. Load Balancing is done with the help of load balancers where each incoming request is redirected and is transparent to client who makes the request. Based on predetermined parameters, such as availability or current load, the load balancer uses various scheduling algorithm to determine which server should handle and forwards the request on to the selected server.

P2P-Based, Multi-Attribute Resource Discovery under Real-World Resources and Queries

Collaborative peer-to-peer (P2P), grid, and cloud computing rely on resource discovery (RD) solutions to aggregate groups of multi-attribute, dynamic, and distributed resources. However, specific characteristics of real-world resources and queries, and their impact on P2P-based RD, are largely unknown. We analyze the characteristics of resources and queries using data from four real-world systems. These characteristics are then used to qualitatively and quantitatively evaluate the fundamental design choices for P2P-based multi-attribute RD. The datasets exhibit several noteworthy features that affect the performance. For example, compared to uniform queries, real-world queries are relatively easier to resolve using unstructured, superpeer, and single-attribute-dominated query-based structured P2P solutions, as queries mostly specify only a small subset of the available attributes and large ranges of attribute values. However, all the solutions are prone to significant load balancing issues, as the resources and queries are highly skewed and correlated. The implications of our findings for improving RD solutions are also discussed.

IJARCCE - Computer and Communication Engineering

Cloud computing is one of the most promising computing field, which has given the new vision to the computing field. Cloud computing has opened a door as a new model for hosting and delivering services over the Internet. The main aim of cloud computing is to provide the resources as a services to the client. The new concept of Federated Cloud Computing in which multiple datacenters are distributed over different regions. Since the evolution of Cloud Computing: load balancing, energy management, VM migration, brokerage policies, cost modelling and security issues are popular research topics in the field. Deployment of real cloud environment for testing or for commercial use is very costly. Cloud simulators help to model various cloud applications and it is very easy to analyse. In this survey, two cloud simulators: CloudSim and CloudAnalyst, with their overview are presented so it can be easily decided which one is suitable for particular research topic. And also the survey on the service broker policy, its issues and available solutions are presented. Because there is always been the requirement to select appropriate datacenter so that further tasks for processing the request should be carried out with efficiency in least response time. So the issue of selecting appropriate datacenter which is known as service broker policy is kind of important.

IJARCCE - Computer and Communication Engineering

2 Abstract: With the initiation of internet in the 1990s to the present day facilities of universal computing, the internet has reformed the computing world in a drastic way. It has travelled from the concept of parallel computing to distributed computing to cluster computing to grid computing to utility computing to virtualization and recently to cloud computing, in future Internet of Things. Virtualization and utility computing can be stated as key concept of cloud. As cloud computing can be specified as a realization of utility computing. Although the idea of cloud computing has been around for quite some time, it is an evolving field of computer science. Since the evolution of cloud computing: Load balancing, energy management, VM migration, server consolidation, cost modelling and security issues are the popular research topic in this field. Deploying real cloud for testing or for commercial use is very costly. Cloud computing model have complex provisioning, composition, configuration, and deployment requirements. Evaluating the performance of Cloud provisioning policies, application workload models, and resources performance models in a repeatable and controllable manner under varying system and user configurations and requirements is difficult to accomplish. To overcome this challenge, cloud simulator is needed. In this paper basic of cloud simulator is discussed, and major focus is on cloudsim- a simulator for management of vm. The CloudSim toolkit supports both system and behaviour modeling of Cloud system components such as data centers, virtual machines (VMs) and resource provisioning policies. It implements generic application provisioning techniques that can be extended with ease and limited effort. Currently, it supports modeling and simulation of Cloud computing environments consisting of both single and inter-networked clouds (federation of clouds). In this paper how cloudsim work, its architectural design, highlighting important features and give brief overview of its functionalities is presented.

Mobile Resource Management Load Balancing Strategy

This paper is dealing with mobile resource management that distributes mobile device processes between cloud computing virtual machine and mobile device. Expectation is that mobile device user experience will increase. Load balancing is an important part of mobile resource management. In order to be able to solve load balancing issues, discussion is made how currently available methods can be adapted to our resource manager. In this article, we investigate load balancing procedures, methods and their customization, with a particular attention on mobile and cloud computing requirements. As a result, we expect that important design aspects will become apparent. Profile based load balancing method is recommended, what combines static and dynamic load balancing strategy. Using profile to identify usage scenario of mobile device can lead to increased system responsiveness, thus experience improvement.

Expert Grid: New Type of Grid to Manage the Human Resources and Study the Effectiveness of Its Task Scheduler

Grid systems are the federation of resources from multiple locations to reach a common goal. In previous researches, the focused resources were only CPU, electronic devices, storage devices, network, and software applications, but till now, human resource (HR) has not been considered as Grid resources in details, despite its importance in many fields of science and society. HR is one of the important assets of organizations and plays a significant role in their success. Common and local methods to share and manage the human’s skills and knowledge in large geographically dispersed organizations are almost centralized, not reconfigurable and are based on the structure of organization; therefore, the optimal trade-off between the HR and job demands is a challenging problem. In addition, these methods did not consider some important parameters such as security issues, load balancing, organizations costs, unnecessary overtime, and employment issues. In order to overcome these defects, in this paper the new Grid architecture, named Expert Grid (EG), is proposed to find, employ, exploit, share, and manage the HR, its skills and knowledge effectively. EG provides infrastructure to increase the customer satisfactions by optimal distributions of requested jobs from customers. Experimental results show that the EG decreases the HR free time, customers delay time, and HR overtime in order to achieve high level of customers satisfaction and better performance of organizations.

A parallel multidomain strategy to compute turbulent flows in fan-stirred closed vessels

This paper presents a parallel multidomain strategy to compute the turbulent flow in a closed vessel stirred by six fans. The method is based on running multiple instances of the same solver, working on different subdomains and communicating through small overlapping zones where interpolations allow to handle moving meshes. First the accuracy of this Multi Instances Solver Coupled on Overlapping Grids (MISCOG) approach is evaluated for the convection of a single vortex. Load balancing issues on parallel machines are discussed and a performance model is proposed to allocate cores to each code instance. Then, the method is applied to the LES of a closed vessel stirred by six fans. Mean and fluctuating fields obtained by the LES are compared to experimental data. Finally, the structure of the turbulence generated at the center of the vessel is studied and the mechanisms allowing turbulence to travel from the fans to the center of the vessel are analyzed.

Built-in Stigmergy-Based Load Balancing Model for HPC Clusters

HPC load balancing aims to manage and share multiple computational resources in order to increase system performance. Several researches have been conducted on the issue of HPC load balancing. Their common point is that they consider processes (workload) as passive units to be balanced among cluster nodes (computational resources) to ensure optimal load balancing. The hypothesis behind our work assumes that the unit to be balanced, i.e. a process (computational task), is more likely to know its needs, and thus more likely to decide whether to migrate to another node or not. Therefore, the load balancing issue is studied at the parallel programming phase. Generated processes, after being recompiled using our LB integrator tool, will be systematically granted a built-in intelligence. The latter enables them to monitor their environment and migrate from one node to another in case of overloading. For this end, we have designed a built-in load balancing model based on stigmergy and collective intelligence concepts (BSLB). A novel P2P cluster computing system has been designed using Repast toolkit and the BSLB algorithm, allowing arbitrary users to initiate intelligent processes.