Search

The emerging network function virtualization is migrating traditional middleboxes, e.g., firewalls, load balancers, proxies, from dedicated hardware to virtual network functions (VNFs) running on commercial servers defined as network points of presence (N-PoPs). VNFs further chain up for more complex network services called service function chains (SFCs). SFCs introduce new flexibility and scalability which greatly reduce expenses and rolling out time of network services. However, chasing the lowest cost may lead to congestion on popular N-PoPs and links, thus resulting in performance degradation or violation of service-level agreements. To address this problem, we propose a novel scheme that reduces the operating cost and controls network congestion at the same time. It does so by placing VNFs and routing flows among them jointly. Given the problem is NP-hard, we design an approximation algorithm named candidate path selection (CPS) with a theoretical performance guarantee. We then consider cases when SFC demands fluctuate frequently. We propose an online candidate path selection (OCPS) algorithm to handle such cases considering the VNF migration cost. OCPS is designed to preserve good performance under various migration costs and prediction errors. Extensive simulation results highlight that CPS and OCPS algorithms perform better than baselines and comparably to the optimal solution.

Datacenter networks (DCNs) play a key role in providing cloud services. The energy consumption and cost of a DCN are growing sharply with the extensions of network bandwidth and network size. The energy consumption, complexity and cost of a DCN depend on some design factors such as the topology structure, addressing scheme and routing mechanism. A novel addressing and routing architecture for cloud-service DCNs with regular topologies is proposed in this paper. First of all, we propose a port-based source-routing addressing (PSRA) scheme, which makes the table-lookup operation unnecessary and decreases the switch complexity. Next, leveraging the characteristics of PSRA and the regularity of DCN topologies, an extremely simple routing mechanism is designed, without switch involvement, control message interaction and topology information storage. Lastly, a high-efficiency fault-tolerance mechanism is proposed for the addressing and routing architecture. The analysis, implementation and simulation results indicate that the proposed architecture not only decreases the energy consumption and thus the cost of a DCN, but also enhances the routing performance and solves the fault-tolerance problem in a very efficient way.

In the Internet age, analyzing the behavior of online users can help webstore owners understand customers’ interests. Insights from such analysis can be used to improve both user experience and website design. A prominent task for online behavior analysis is clickstream mining, which consists of identifying customer browsing patterns that reveal how users interact with websites. Recently, this task was extended to consider weights to find more impactful patterns. However, most algorithms for mining weighted clickstream patterns are serial algorithms, which are sequentially executed from the start to the end on one running thread. In real life, data is often very large, and serial algorithms can have long runtimes as they do not fully take advantage of the parallelism capabilities of modern multi-core CPUs. To address this limitation, this paper presents two parallel algorithms named DPCompact-SPADE ( D epth load balancing P arallel Compact-SPADE) and APCompact-SPADE ( A daptive P arallel Compact-SPADE) for weighted clickstream pattern mining. Experiments on various datasets show that the proposed parallel algorithm is efficient, and outperforms state-of-the-art serial algorithms in terms of runtime, memory consumption, and scalability.

The execution time of a stage is extended by a few slow running tasks in Spark computing environments. Due to the characteristics of tasks and the complexity of runtime environments, the Spark original speculative execution strategy and its improved versions cannot deal with this problem effectively. In this paper, we propose a novel strategy called ETWR to improve the efficiency of speculative execution in Spark. We consider the heterogeneous environments when we around to tackle the three key points of speculative execution: straggler identification, backup node selection and effectiveness guarantee. Firstly, we divide the task into sub-phases and use both the process speed and progress rate within a phase to find the straggler promptly. Secondly, we use the Locally Weighted Regression model to estimate the execution time, which will be used to calculate the task remaining time and backup time. Thirdly, we present iMCP model to guarantee the effectiveness of speculative tasks, which can additionally keep load balancing for nodes. Finally, the factors of fast node and better location are considered when selecting proper backup nodes. Extensive experiments show that ETWR can reduce the job execution time by 23.8%, and improve the cluster throughput by 33.2% compared with Spark-2.2.0.

We consider a large-scale service system where incoming tasks have to be instantaneously dispatched to one out of many parallel server pools. The user-perceived performance degrades with the number of concurrent tasks and the dispatcher aims at maximizing the overall quality-of-service by balancing the load through a simple threshold policy. We demonstrate that such a policy is optimal on the fluid and diffusion scales, while only involving a small communication overhead, which is crucial for large-scale deployments. In order to set the threshold optimally, it is important, however, to learn the load of the system, which may be unknown. For that purpose, we design a control rule for tuning the threshold in an online manner. We derive conditions which guarantee that this adaptive threshold settles at the optimal value, along with estimates for the time until this happens. In addition, we provide numerical experiments which support the theoretical results and further indicate that our policy copes effectively with time-varying demand patterns., Comment: 51 pages, 6 figures

Program execution on a multicore platform, even a homogeneous one, does not guarantee performance gains. The key is in letting each of the cores contribute to the overall computation for the maximum amount of time. For anything but trivial problem settings, and especially for heterogeneous platforms like ones made of CPU and GPU cores, this calls for some form of load balancing, i.e., explicit or implicit workload shifts between the cores, so that idle times are eliminated to the maximum degree possible. In this chapter we examine the two generic categories of load balancing algorithms: static which can be considered pro-active, meaning that they pre-calculate load assignments without expensive communications taking place, and dynamic which can be considered re-active as they can adapt to changes in the execution environment. The chapter is concluded by two cases studies that show how both static and dynamic load balancing can be used to evenly distribute the work between the nodes of a heterogeneous platform, that combines both CPU and GPU nodes.

In this paper, static wireless network load balance algorithm is proposed, that use only optimal paths from point-to-point to achieve good load balance. This algorithm based on Enhanced version of Elman Recurrent Neural Network (EERNN) to make load balance decision depended on two metrics (traffic load on node and probability of link failure). This algorithm make good work in terms of both metrics simultaneously. Also, this algorithm use only local information. The execution of the proposed algorithm is compared with ERNN based on traffic load on node and probability of link failure. © 2018 JASET, International Scholars and Researchers Association Author Biographies Lubna Najah Rasoul Mazaya University Collage Wijdan Rasheed Abd Al-Hussain University of Thi-Qar

Existing synchronous network search data extraction techniques suffer from load unbalance in the face of increasing number of concurrent users. For this reason, this paper presents the research on the optimum synchronous network search data extraction based on swarm intelligence algorithm. A traffic balancing model is constructed to determine the node balance state according to the traffic variation and influence factors of network nodes. Under the condition that the node state is known, the swarm intelligence algorithm is used to cluster the data to be synchronized and adjust the node state so that it is kept stable throughout the synchronization process. The clustered data act as the target to connect the user with the server side to achieve the optimum network search data extraction and synchronization. The experimental results show that when the number of concurrent network users is increasing, the designed technique features stable load balancing, and achieves optimum data extraction performance and low execution cost when the task completion time is less than 0.5 s.

Background: Load balancing of communication-intensive applications, allowing efficient resource utilization and minimization of power consumption is a challenging multi-objective virtual machine (VM) placement problem. The communication among inter-dependent VMs, raises network traffic, hampers cloud client’s experience and degrades overall performance, by saturating the network. Introduction: Cloud computing has become an indispensable part of Information Technology (IT), which supports the backbone of digitization throughout the world. It provides shared pool of IT resources, which are: always on, accessible from anywhere, at anytime and delivered on demand, as a service. The scalability and pay-per-use benefits of cloud computing has driven the entire world towards on-demand IT services that facilitates increased usage of virtualized resources. The rapid growth in the demands of cloud resources has amplified the network traffic in and out of the datacenter. Cisco Global Cloud Index predicts that by the year 2021, the network traffic among the devices within the datacenter will grow at Compound Annual Growth Rate (CAGR) of 23.4% Methods: To address these issues, a communication cost aware and resource efficient load balancing (CARE-LB) framework is presented, that minimizes communication cost, power consumption and maximize resource utilization. To reduce the communication cost, VMs with high affinity and inter-dependency are intentionally placed closer to each other. The VM placement is carried out by applying the proposed integration of Particle Swarm Optimization and non-dominated sorting based Genetic Algorithm i.e. PSOGA algorithm encoding VM allocation as particles as well as chromosomes. Results: The performance of proposed framework is evaluated by the execution of numerous experiments in the simulated datacenter environment and it is compared with the state-of-the-art methods like, Genetic Algorithm, First-Fit, Random-Fit and Best-Fit heuristic algorithms. The experimental outcome reveals that the CARE-LB framework improves 11% resource utilization, minimize 4.4% power consumption, 20.3% communication cost with reduction of execution time up to 49.7% over Genetic Algorithm based Load Balancing framework. Conclusion: The proposed CARE-LB framework provides promising solution for faster execution of data-intensive applications with improved resource utilization and reduced power consumption. Discussion: In the observed simulation, we analyze all the three objectives, after execution of the proposed multi-objective VM allocations and results are shown in Table 4. To choose the number of users for analysis of communication cost, the experiments are conducted with different number of users. For instance, for 100 VMs we choose 10, 20,...,80 users, and their request for VMs (number of VMs and type of VMs) are generated randomly, such that the total number of requested VMs do not exceed number of available VMs.

With the rapidly growing traffic demand, the cellular industry is moving toward heterogeneity to fulfill the heavy traffic requirement. The existing heterogeneous networks are comprised of the Long-Term Evolution (LTE), LTE-Advanced (LTE-A), and other compatible access technologies. It has been identified that due to the shortage of capacity, cell performance can deteriorate. To deal with users’ requirements, the 5G heterogeneous network architectures comprising LTE-advanced access technology with the combination of comparatively low configured small base stations, and macro eNodeBs (MeNB) have been extended as a solution. Therefore, operators have introduced the 5G with the existing LTE Advanced Heterogeneous Network (5GLHN) where small cells such as Home eNodeBs (HeNBs) are deployed overlapping with the conventional macro eNodeB (MeNB). However, in highly dense and closely compacted 5GLHNs, the cell capacity can still be lower than what is on-demand, thus affecting the system throughput. This article proposes a framework to maximize the throughput in 5GLHNs through a Constriction Factor Particle Swarm Optimization (CFPSO) technique of cell association and load-balancing algorithm to enhance the throughput of the 5GLHN. The proposed approach is designed to offload the traffic of MeNB Users (MUEs) to the small cells (HeNBs). The convergence, cumulative distribution function of the UEs rate, average throughput, and allocation time are analyzed to evaluate the performance of the proposed CFPSO approach. Simulation results reveal that the throughput of the proposed approach is improved by up to 44.08% of the existing index-based approach and by 94.20 \% of the existing Matching with Minimum Quota (MMQ) approach.

To meet the ever-increasing demands for high-quality network services, a software defined network (SDN) can support various virtual network functions (VNFs) using virtualization technology. Due to network dynamics, an SDN needs to be updated frequently to optimize various performance objectives, such as load balancing. Most previous solutions first determine a new network configuration (e.g., target VNF placement and flow routing) based on the current workload, and then update the VNF placement and routing paths of the existing flows. However, due to massive VNF's state migration and slow update of the flow table, unacceptable update delay may occur, especially in large or frequently changed networks. In this paper, we address the real-time network update, which jointly considers the optimization of the service function chain (SFC) update and the routing update. We propose the delay-satisfied NFV-enabled network update (DSNU) problem, and prove its NP-Hardness. We design an algorithm with bounded approximation factor to solve this problem. To further reduce the delay, we also design an efficient algorithm for the update scheduling. The experimental results show that our method can reduce the network update delay by about 86% compared with the previous network update methods while preserving a similar network performance, \ie, the VNF instance load ratio increases by less than 5%.

As battery technology has not advanced as quickly as other technologies like semiconductor technology, energy efficiency is still one of the primary concerns in Wireless Sensor Networks (WSNs). One of the finest Wireless Energy Transfer (WET) technique is Magnetic Resonant Coupling (MRC) used to re-energize the nodes in WSNs. Recent advancements in MRC show that more than one node can be re-charged at the same time. This paper aims to exploit multiple nodes WET to overcome the issues pertaining to energy in Wireless Sensor Network (WSN). For transferring energy to various static sensor nodes deployed in a network, Unmanned Aerial Vehicle (UAV) is considered which travels inside a network and performs a process of re-charging. Further, WSNs being most acceptable face issues like: battery constraint, and congestion. Therefore, in order to address such issues, we have proposed MRCRLB (Magnetic Resonant Coupling Based Recharging and Load Balancing) scheme. Then, finally the proposed model is tested on various modulation schemes at the physical layer of WSNs to find out most appropriate modulation scheme type for lifetime extension model (MRCRLB).

Macro base stations are overlaid by small cells to satisfy the demands of user equipment in heterogeneous networks. To provide wide coverage, some small cells are not directly connected to macro base stations and thus backhaul connections are required to connect small cells to macro base stations. Millimeter wave backhauls which have high bandwidths are preferred for small cell backhaul communication, since they can increase the capacity of network considerably. In this context, association of user equipment to base stations becomes challenging due to the backhaul architecture. Considering environmental concerns, energy efficiency is a vital criterion in designing user association algorithms. In this paper, we study the user association problem aiming at the maximization of energy efficiency given a specific spectral efficiency target. We develop centralized and distributed user association algorithms based on sequentially minimizing the power consumption. Finally, we evaluate the performance of the proposed algorithms under two scenarios and show that they achieve higher energy efficiency compared to the existing algorithms in the literature, while maintaining high spectral efficiency and backhaul load balancing.

Traffic management in software-defined networks (SDNs) is critical for efficient bandwidth utilization and resource provisioning. Recent works on SDN load balancing (LB) have focused on identifying and rerouting elephant flows (EFs) for effective bandwidth usage. These techniques have some limitations, such as using source-to-destination hop count as the primary rerouting metric and not differentiating the types of flow that result in frequent resource conflicts when handling EF with long-lived bandwidth. Besides, current EF detection techniques use predefined bandwidth-use thresholds that cannot adapt to the ever-changing traffic condition. Also, detecting EF on switches results in high controller-switch bandwidth and high EF detection time. This study presents an ant colony optimization-based technique for rerouting EFs while considering load-balancing in the SDN links. This technique, called DPLBAnt, is formulated as a shortest-path problem in SDN that can alleviate the high controller-switch load. The proposed technique first detects EF by using a pair of classifiers on both SDN controller and switches. Most EF candidates are sifted on the switches, resulting in accurate and efficient detection of EF. Then, DPLBAnt obtains the global state of the SDN from which the most optimal paths for congested links are retrieved, and EF are redirected accordingly. The performance of the proposed DPLBAnt has been extensively simulated. Results indicate its superior performance over Equal-Cost Multi-Path (ECMP) and FlowSeer techniques in terms of average end-to-end delay (54% and 7.9% better), average network throughput ( 3 . 5 × and 1 . 5 × better), and average packet loss (18% and 10% better) respectively. The overall performance indicates that the proposed LB technique based on detection and rerouting of EFs can improve SDN’s overall performance.

Data center networks typically adopt multi-rooted tree topologies to provide high bisection bandwidth. Various fine-grained load balancing schemes have been proposed to split flows across multiple paths. However, data center networks suffer from many uncertainties such as highly dynamic traffic. These uncertainties easily make network become asymmetric, resulting in significant packet reordering. Unfortunately, existing solutions passively deal with packet reordering based on a threshold and hardly adapt to asymmetric networks because of lacking the explicit reordering feedback. These solutions either fail to quickly respond to packet loss or cause unnecessary fast retransmission, which reduces link utilization and increases flow completion time. In this paper, we propose a fine-grained load balancing scheme RMC to eliminate the impact of packet reordering and handle uncertainties in asymmetric networks. To avoid unnecessary fast retransmission, the switch proactively identifies reordered packet according to local queue length and global path latency. Furthermore, we employ a coding technique with redundancy optimization to reduce long-tailed flow completion time under network asymmetry. Through a series of large-scale NS2 simulations and testbed experiments, we demonstrate that RMC effectively avoids unnecessary fast retransmission under different network scenarios and reduces flow completion time by up to 72% compared with state-of-the-art schemes.

The introduction of artificial intelligence (AI) into edge computing could significantly improve its quality of service. Connecting them into a system can provide services for a wider range. However, due to the mobility of the crowd and mobile devices, the load imbalance issue of these interconnected intelligent edge servers (IESs) will cause severe impacts on their service performance. To this end, we investigate load balancing for the distributed IESs from the game theoretic perspective and propose a state-based distributed learning algorithm. Firstly, by modelling the IES cost as the deviation between its execution time and the system average execution time, we formulate load balancing as a state-based game where each IES competes to maximize its own utility. Secondly, according to the definition of the recurrent state Nash equilibrium, we prove that this game has such an equilibrium by establishing a potential function at each reachable state. Finally, we propose a state-based distributed learning algorithm to obtain the pure Nash equilibrium strategy of each IES. Then, an ordinary differential equation is derived to prove the convergence of the algorithm. In comparison with existing works, our approach could largely improve load balancing for the distributed IESs and thus enhance their service performance.

Wireless Sensor Networks is one of the most significant area of research where large number of sensor nodes that are distributed in a geographical area operate on limited battery power. As these networks, depending on the application, are sometimes deployed in in-hostile environment, which makes energy conservation one of the major challenge faced in WSN. To reduce energy consumption, clustering is considered to be the most efficient technique. This work proposes a new clustering approach that decreases energy consumption and results in prolonged network lifetime which is an important requirement for networks operating in inaccessible areas. In the proposed approach, heterogeneity is also implemented to increase the stability and energy efficiency of random networks. We have evaluated the efficiency of the proposed protocol through simulations and comparison is performed with well-known existing distributed protocols. Proposed approach shows efficient results in terms of the stability period for different network configurations and Base Station positions. Also, the results are found better in terms of number of alive nodes and network lifetime.

Perancangan jaringan yang baik diperlukan untuk menghasilkan jaringan yang optimal karena kebutuhan akses internet yang terus meningkat. Permasalahan yang biasa terjadi adalah terkait ketersediaan jaringan. Solusi untuk meminimalkan permasalahan tersebut yaitu dengan menerapkan Failover dan Load Balancing. Konsep yang digunakan adalah Failover dan Load Balancing dengan metode Per Connection Classifier (PCC). Load Balancing adalah teknik untuk pendistribusian traffic dengan dua jalur gateway atau lebih sehingga traffic dapat berjalan lebih optimal. Metode yang digunakan adalah Per Connection Classifier yang dapat mengelompokkan koneksi menjadi beberapa kelompok. Selanjutnya, teknik Failover juga diterapkan untuk fungsi redundancy link. Selanjutnya dilakukan pengujian pendistribusian beban traffic, pengujian Failover dan analisis perbandingan QoS dengan parameter jitter, packet loss, throughput, dan delay terhadap kinerja Load Balancing metode PCC berdasarkan Classifier Both Addresses, Both Ports, dan Both Addresses and Ports. Hasil menunjukkan bahwa sistem dapat mendistribusikan beban traffic dan dapat memberikan fungsi redundancy, sedangkan untuk pengujian QoS secara keseluruhan, Classifier Both Addresses mendapatkan hasil kinerja yang lebih bagus dibandingkan classifier lainnya.

Wireless sensor networks are usually deployed in areas with relatively harsh natural environments, and the collection node transmits data to the destination node through a multi-hop route. Therefore, how to effectively plan the transmission path is an important issue. This paper combines the unbiased gray model with the Markov chain model to establish an unbiased gray Markov chain model, and points out that the unbiased gray Markov chain model also has shortcomings in parameter selection. The particle swarm algorithm is used to improve it, and the mathematical model, calculation principle and various parameters of the particle swarm optimization algorithm are introduced, and the implementation flow chart of the particle swarm algorithm is given. Aiming at the shortcomings of the unbiased gray Markov chain model, the particle swarm algorithm and the unbiased gray Markov chain model are combined to form the particle swarm unbiased gray Markov chain model. The simulation environment and the training environment of the particle swarm unbiased grey Markov chain model were designed in the experiment. The node scheduling optimization experiment proves that the scheduling method based on the particle swarm unbiased grey Markov chain model has achieved better results in coverage and energy consumption balance than the random and shortest distance method. In the routing experiment, the experimental analysis of the node’s Q value proved the convergence of the algorithm, and compared with other protocols, it proved that the routing algorithm can effectively extend the network life cycle and achieve load balancing.

Handling the ever-increasing complexity of mesh generation codes along with the intricacies of newer hardware often results in codes that are both difficult to comprehend and maintain. Different facets of codes such as thread management and load balancing are often intertwined, resulting in efficient but highly complex software. In this work, we present a framework which aids in establishing a core principle, deemed separation of concerns, where functionality is separated from performance aspects of various mesh operations. In particular, thread management and scheduling decisions are elevated into a generic and reusable tasking framework. The results indicate that our approach can successfully abstract the load balancing aspects of two case studies, while providing access to a plethora of different execution back-ends. One would expect, this new flexibility to lead to some additional cost. However, for the configurations studied in this work, we observed up to $$13\%$$ speedup for some meshing operations and up to $$5.8\%$$ speedup over the entire application runtime compared to hand-optimized code. Moreover, we show that by using different task creation strategies, the overhead compared to straight-forward task execution models can be improved dramatically by as much as $$1200\%$$ without compromises in portability and functionality.

Cloud computing provides computing resources on demand of users without their direct management. In this cloud paradigm, scheduling the tasks and allocating the resources become major aspect for cloud infrastructure as a service (IaaS). There are more existing algorithms and techniques suggested for task allocation problem. Still there is challenging research on efficient scheduling. To address this issue, many researches are in progress and all of them having their own drawbacks. In this paper, we are proposing a queue-based scheduling strategy with load balancing called as IQSLB and an extended IQSLB also proposed for dealing with critical situations. The proposed strategy calculates the placement value of the tasks in queue with the current status of the virtual machine (VM) in cluster and reshuffles the task accordingly. The extended IQSLB deals with handling deadlock situation where VM cannot adopt task for execution and task will be reshuffled with another task in another queue. The proposed strategy is compared with few existing systems, and the performance evaluation proves that IQSLB schedules tasks more efficiently than other systems. Our proposed IQSLB takes 75 s to allocate 1000 tasks by using 55 virtual machines which is much lesser than existing techniques.

Over the years cloud computing has seen significant evolution in terms of improvement in infrastructure and resource provisioning. However the continuous emergence of new applications such as the Internet of Things (IoTs) with thousands of users put a significant load on cloud infrastructure. Load balancing of resource allocation in cloud-oriented IoT is a critical factor that has a significant impact on the smooth operation of cloud services and customer satisfaction. Several load balancing strategies for cloud environment have been proposed in the past. However the existing approaches mostly consider only a few parameters and ignore many critical factors having a pivotal role in load balancing leading to less optimized resource allocation. Load balancing is a challenging problem and therefore the research community has recently focused towards employing machine learning-based metaheuristic approaches for load balancing in the cloud. In this paper we propose a metaheuristics-based scheme Data Format Classification using Support Vector Machine (DFC-SVM), to deal with the load balancing problem. The proposed scheme aims to reduce the online load balancing complexity by offline-based pre-classification of raw-data from diverse sources (such as IoT) into different formats e.g. text images media etc. SVM is utilized to classify “n” types of data formats featuring audio video text digital images and maps etc. A one-to-many classification approach has been developed so that data formats from the cloud are initially classified into their respective classes and assigned to virtual machines through the proposed modified version of Particle Swarm Optimization (PSO) which schedules the data of a particular class efficiently. The experimental results compared with the baselines have shown a significant improvement in the performance of the proposed approach. Overall an average of 94% classification accuracy is achieved along with 11.82% less energy 16% less response time and 16.08% fewer SLA violations are observed.

In order to improve the effect of intelligent terminal container management, this paper improves the Internet of Things and big data technology, analyzes the RFID middleware architecture based on the actual needs of container handling management, and proposes a new method of RFID middleware load balancing. Moreover, this paper combines the Internet of Things technology and big data technology to analyze the terminal container loading and unloading process and build a corresponding intelligent system. After constructing a terminal container handling intelligent management system based on the Internet of Things and big data technology, the performance of the system is verified, and multiple sets of simulation data are used to conduct research. The experimental research results show that the terminal container handling management system based on the Internet of Things and big data constructed in this paper basically meets the actual needs of use.

Nowadays, video conferencing is one of the state-of-the-art and highly demanded technologies for conducting online communications. Providing real-time and rapid connection establishment is one of the chief principles of video conferences which is competently achieved by Multi-Protocol Label Switching (MPLS). Also, the emerging concept of Next Generation Networks (NGN) has accelerated the advancement of IP-based multimedia networks including Voice over IP (VoIP), Video on Demand (VoD) and IPTVs. MPLS networks have formed the foundation of fulfilling multimedia requirements in NGN. Therefore, Label Switching Path (LSP) routing is one of the highlighted challenges of Traffic Engineering (TE) in MPLS networks. MPLS routing algorithms attempt to increase the acceptance rate of requests and consequently meet the Quality of Service (QoS) constraints satisfaction. In this paper, we propose a novel routing algorithm that applies fuzzy rules to meet the bandwidth and end-to-end delay constraints in the routing. The proposed fuzzy mechanism constitutes a predicting system based on weighted fuzzy rules that filters the requests with higher resource demands. We name the proposed method as Fuzzy Bandwidth and Delay guaranteed Routing Algorithm (FBDRA). Schematically, FBDRA aims to defer the requests with maximum bandwidth and minimum end-to-end delays. All simulation experiments are carried out in MATLAB R2019a on different scenarios. We have gauged several metrics such as the number of accepted requests, average path length, energy consumption, load balancing, and average end-to-end delay to evaluate our proposed algorithm performance. The simulation results evidence the superiority of FBDRA in video conference applications.

One of the most important means to position abnormal devices is to efficiently utilize the resources of wireless sensor network (WSN) and make proper analysis of the relevant data. Therefore, this paper constructs an indoor positioning and prewarning system that utilizes energy efficiently and achieves a long lifecycle. Firstly, the adjacent round iteration load balancing (ARILB) routing algorithm was proposed, which elects the cluster heads (CHs) by the adjacent round strategy. In this way, the random components were eliminated in CH election. Next, a short-distance multifrequency routing strategy was constructed between CHs to transmit the information to the sink, and a positioning algorithm was designed called ARILB-received signal strength (RSS). The ARILB-RSS positioning algorithm traverses the triangles formed by anchor nodes, forming multiple sets of ranging points; then, the optimal anchor node is recorded, and the path loss factor is iterated to reduce the positioning error. Simulation shows that the network survives 54.5% longer using ARILB than using the distributed energy-efficient clustering (DEEC) algorithm; the packet delivery rate using ARILB was about 139% higher than that of low energy adaptive clustering hierarchy (LEACH) algorithm and 35% higher than that of uneven clustering routing algorithm based on chain-cluster type (URCC) algorithm; ARILB-RSS reduced the ranging error by 14.31% and then the positioning error by 26.79%.

As an important part of the Internet of Things, mobile wireless sensor network (MWSN) will generate traffic with different service quality requirements due to the multi-sensor integration of nodes and application diversity. Due to the topological changes, resource constraints and self-organizing characteristics of MWSN, the Per-Hop Behavior (PHB) approach in the traditional Diff-Serv model has many challenges in providing differential service. In this paper, a path reservation multipath routing (PRMR) protocol is proposed, which can provide a suitable path for each type of traffic with service requirements. PRMR protocol includes path discovery algorithm and packet scheduling algorithm. In addition, the path scheduling model and scheduling algorithm in PRMR solve the problem of load imbalance among reserved paths caused by different types of traffic. In scenarios with different number of nodes, three types of traffic, integrity sensitive data, delay sensitive data and normal data, are used to verify the performance of PRMR differential service and load balancing. Simulation experiments not only compare the quality of service provided by each reserved path in PRMR, but also compare the differential service performance between PRMR and several multipath routing algorithms. Simulation results show that PRMR routing algorithm can guarantee high packet delivery rate and low delay for integrity-sensitive data and delay-sensitive data, respectively. In addition, the simulation results of the average residual energy index show that the protocol can achieve network energy balance.

5G micro infrastructure comprising micro and picocells would play a pivotal role in densifying the network to provide ample coverage. However, a disproportional association of mobile devices with these small cells would cause hotspots and load imbalance. In such a network, a few micro or picocells suffer from network congestion. While many others are underutilized, experience lower throughput, and operate below the potential network capacity. To mitigate this drawback, some means of Load Balancing (LB) would be essential in heterogeneous and homogenous networks. To achieve this, we propose an extreme Swap-based Load Balancing (SLB) algorithm between APs, which minimizes the load imbalance at cell edges. The experimental setup uses a dataset contributed by Irish mobile operators. Our results reveal SLB with biasing reduces the load imbalance by a factor of 7.14% compared to the optimal uni-transfer algorithm. Against other state-of-the-art algorithms, it betters by 22.24%. SLB with biasing delivers both lesser load imbalance in APs and signal quality amongst users.

This manuscript presents design and analysis of a set of reference-tracking control policies for large-scale coordination of distributed energy resources (DERs) and quantifies tracking errors that arise due to heterogeneity in the power ratings for a fleet of DERs. In particular, the relay-based, reference-tracking control strategy that underpins much of packetized energy management (PEM) is augmented to uniquely leverage PEM’s energy packet request mechanism to optimize the number of accepted requests and to explicitly consider the quality of service (QoS). In addition, tracking errors from modeling a heterogeneous fleet of packetized DERs with a group of homogeneous macromodels are analytically derived for relevant PEM information scenarios. Finally, simulation-based analysis validates the results and shows that PEM is suitable for providing load balancing and ramping services for the grid.

A challenge that heterogeneous system programmers face is leveraging the performance of all the devices that integrate the system. This paper presents Sigmoid, a new load balancing algorithm that efficiently co-executes a single OpenCL data-parallel kernel on all the devices of heterogeneous systems. Sigmoid splits the workload proportionally to the capabilities of the devices, drastically reducing response time and energy consumption. It is designed around several features; it is dynamic, adaptive, guided and effortless, as it does not require the user to give any parameter, adapting to the behaviourof each kernel at runtime. To evaluate Sigmoid's performance, it has been implemented in Maat, a system abstraction library. Experimental results with different kernel types show that Sigmoid exhibits excellent performance, reaching a utilization of 90%, together with energy savings up to 20%, always reducing programming effort compared to OpenCL, and facilitating the portability to other heterogeneous machines. This work has been supported by the Spanish Science and Technology Commission under contract PID2019-105660RB-C22 and the European HiPEAC Network of Excellence.

Virtual machine migration (VMM) is a crucial way to ensure load balancing and save energy consumption of cloud hosts. In most previous studies, they apply the same virtual machine migration strategy for overloaded hosts, which ignore the possible communication between the migrated VM, and other factors. Thus, they are prone to incur excessive network overhead. This paper proposes a virtual machine migration strategy based on the three-way decision (VMM-3WD) to save cloud hosts’ energy consumption while considering the network correlation between virtual machines. The strategy first is to classify hosts into overloaded hosts, regular load hosts, and under-loaded hosts based on their load situation. Then, different migration strategies are targeted developed for these three types of cloud hosts. Specifically, the approach migrates the VMs in under-loaded hosts to regular load hosts. And then, the approach further develops two thresholds to divides overloaded hosts into massively overloaded hosts, moderately overloaded, and lightly overloaded hosts. The migration decision of VMs in all stages pursuing the goal of reducing the energy consumption of the network during the migration process. The experimental results show that the proposed algorithm can reduce the energy consumption generated by virtual machine migration and data center (DC) energy consumption while meeting the SLA.

Mobile edge computing (MEC) is a key feature of next-generation heterogeneous networks aimed at providing a variety of services for different applications by performing related processing tasks closer to the user equipment. In this research, we investigated on connection management approaches in multi-access edge computing systems. This paper presents joint radio resource allocation and MEC optimization in a multi-layer NOMA HetNet in order to maximize system’s energy efficiency. The continues carrier allocation and handoff decision variables, in addition to the interference incorporated in the goal function, modifies the primary optimization problem to a mixed integer nonlinear programming. Network selection is done statically based on the Analytic Hierarchy Process, and station selection is done dynamically based on the Data Envelope Analysis method. Also, an effective feedback mechanism has been designed in collaboration with the server resource manager to solve a global optimization problem in order to load balancing and meet the users quality of service constraints simultaneously. To reduce the computational complexity and to achieve a locally optimal solution, we applied variable relaxation and majorization minimization approach in which offloading decision and multi-part Markov noise analysis have been developed to model users’ preferences without the need for explicit information from the users. Based on the simulations, the proposed approach not only results in a significant increase of system’s energy efficiency but also enhances the system throughput in multiple-source scenarios.

With the development of social economy, English is still the main language of global communication, and China’s demand for English translators is also higher and higher. In college education, English translation is a very popular and important subject. Although the teaching mode is constantly changing, it still lags behind the speed of social and economic development. Traditional English translation teaching has many disadvantages in teaching theory and methods. This paper discusses the problems existing in translation teaching for English majors and then puts forward a new teaching mode of educational multiattitude resources, to really improve the quality of translation teaching for English majors. The resource pool of cloud computing is composed of a large number of resource nodes with different performances, but as the number of users gradually grows, the demand also increases. How to efficiently allocate large-scale cloud tasks to limited resource nodes and to achieve load balancing is an important issue that needs to be studied in cloud computing. This paper first uses the literature method, in-depth study of the theoretical development, and application bottleneck of education multiposture teaching mode at the same time to understand the problems existing in the current English translation teaching and design a new teaching mode for experimental application research and finally analyzes the comparison of the English translation ability of the control group and the experimental group without teaching method. After a period of experiment, the average score of the control group is 70.53, and the average score of the experimental group is 75.89. It can be seen that the average score of the experimental group is about 5 points higher than that of the control group; this shows that the intelligent optimization algorithm of cloud computing is combined with translation teaching that can improve the students’ overall English translation level to a certain extent and is more conducive to the students’ comprehensive mastery of knowledge.

Background: The field of cloud computing has been evolving for over a decade now. Load balancing is an important component of cloud computing. Load balancing implies scheduling of cloudlets (tasks) on virtual machines. Since this is an NP-hard problem, various heuristics for load balancing have been proposed in the research literature. The heuristics have been categorized, simulated and benchmarked in various ways; however, the information is scattered across many review articles. Objective: This review aims to bring a broad range of load balancing heuristics found in the research literature under one umbrella. It includes a comprehensive list of heuristics, a holistic set of criteria for their classification, and some key performance metrics and simulation tools used for their benchmarking. An illustration of a fair and comprehensive comparison of heuristics is provided using CloudSim Plus, a recent and advanced simulation tool. Method: The simulations performed with CloudSim Plus employ a generic model of task and machine heterogeneity with Poisson arrival of cloudlets and exponential distribution of cloudlet length to emulate actual cloud-computing scenarios. The simulation results in terms of key performance metrics are used to compare four centralized load balancing heuristics including Join Shortest Queue (JSQ), Join Idle Queue (JIQ), Round Robin and Minimum Completion Time (MCT).

Due to the non-uniformity of the sparse matrix, the calculation of SPMV (sparse matrix vector multiplication) will lead to redundancy in calculation, redundancy in storage, unbalanced load and low GPU utilization. In this study, a new matrix compression method based on CSR and COO is proposed for the above analysis: PBC algorithm. This method considers the load balancing condition in the calculation process of SPMV, and blocks are divided according to the strategy of row main order to ensure the minimum standard deviation between each block, aiming to satisfy the maximum similarity in the number of nonzero elements between each block. This paper preprocesses the original matrix based on block splitting algorithm to meet the conditions of load balancing for each block stored in the form of CSR and COO. Finally, the experimental results show that the time of SPMV preprocessing is within the acceptable range of the algorithm. Compared with the serial code without CSR optimization, the parallel method in this paper has an acceleration ratio of 178x. In addition, compared with the serial code for CSR optimization, the parallel method in this paper has an acceleration ratio of 6x. And a representative matrix compression method is also selected for performing comparative analysis. The experimental results show that the PBC algorithm has a good efficiency improvement compared with the comparison algorithm.

Cloud technology is a utility where different hardware and software resources are accessed on pay-per-user ground base. Most of these resources are available in virtualized form and virtual machine (VM) is one of the main elements of visualization. However, the tasks send by user to cloud may cause the VM to be under loaded or overloaded due to tasks allocation system in VM which lead to the failure of the system or delay the user tasks. Therefore, we propose an improved load balancing technique known as hybridizing artificial bee colony algorithm with Bat algorithm (HBAC). For searching food source employed bee’s use they share the information about to the food source to onlooker bee. In the initialization section equal number of employed bees and onlooker bees used for searching process with the same updation rule which make trapping in search process. Therefore for employed bee the Bat updation rule use in initialization section. When the employed bees share the information with onlooker bee with the help of dancing now it time for onlooker bee to prepare the candidate bee for searching process. Onlooker bees start searching for candidate bee using as technique in this technique it take cycle for searching bee if some tasks are missing in this cycle it take more cycle up to when all tasks are cover in the searching process. This technique take more time for that reason a new technique used in onlooker searching section which make the tasks are into equal part then start searching which was affective and take less time as compare to the previous one. The third modification took place at fitness value of artificial bee colony algorithm where the tasks distribution take more time due to overlapping which affect the tasks accuracy system. The proposed HBAC algorithm was tested and compared with other state-of-the-art algorithms on 200–2000 even tasks by using CloudSim on standard workload format (SWF) data sets file size (200 kb and 400 kb). The proposed HBAC showed an improved accuracy rate in tasks distribution and reduced the makespan of VM in a cloud data center. Based on the ANOVA comparison test results, a 1.25% improvement on accuracy and 0.98% reduced makespan on tasks allocation system of VM in cloud computing is observed with the proposed HBAC.

In server clusters, the static scheduling algorithm has superior performance when the user visits are relatively stable. In the face of sudden increase in user traffic, the dynamic scheduling algorithm has a better load balancing effect than the static scheduling algorithm. However, in the face of complex network environments, the static scheduling algorithm cannot adjust the load according to the performance of the server in real time. The dynamic scheduling algorithm using a single weight to evaluate server performance is unreliable, and load scheduling with reference to the number of connections has uncertainty. In view of this problem, this paper proposes a cluster load balancing algorithm based on dynamic consistent hash based on the study of load balancing based on LVS clusters. By analyzing the performance and load parameters, we divide the request process into in-cycle and out-of-cycle. By setting up the LVS cluster system, the performance weights, load parameters, number of virtual nodes and cycles of the algorithm in this paper are determined experimentally. Finally, the response time and throughput of the algorithm in this paper are compared with the WRR algorithm and WLC algorithm. The results show that the time and throughput of this algorithm are better than WRR algorithm and WLC algorithm.

The growth of cloud application services delivered through data centers with varying traffic demands unveils limitations of traditional load balancing methods. Aiming to attend evolving scenarios and improve the overall network performance, this paper proposes a load balancing method based on an Artificial Neural Network (ANN) in the context of Knowledge-Defined Networking (KDN). KDN seeks to leverage Artificial Intelligence (AI) techniques for the control and operation of computer networks. KDN extends Software-Defined Networking (SDN) with advanced telemetry and network analytics introducing a so-called Knowledge Plane. The ANN is capable of predicting the network performance according to traffic parameters paths. The method includes training the ANN model to choose the path with least load. The experimental results show that the performance of the KDN-based data center has been greatly improved.

In recent years, the application of intelligent transportation systems has gradually made the transportation industry safe, efficient, and environmentally friendly and has led to a broader research prospect of vehicle wireless communication technology. Distributed vehicular self-organizing networks are mobile self-organizing networks in realistic traffic situations. Data interaction and transmission between nodes are achieved through the establishment of a vehicular self-organizing network. In this paper, a multipath routing protocol considering path stability and load balancing is proposed to address the shortcomings of existing distributed vehicular wireless self-organizing routing protocols. This protocol establishes three loop-free paths in the route discovery phase and uses the path stability parameter and load level parameter together to measure the total transmission cost. The one with the lowest total transmission cost is selected as the highest priority path for data transmission in the route selection phase, and the other two are used as alternate paths, and when the primary path breaks, the higher priority of the remaining path will continue to transmit data as the primary route. In this paper, to improve the content distribution performance of target vehicles in scenarios where communication blind zones exist between adjacent roadside units, an assisted download distribution mechanism for video-like large file content is designed in the V2V and V2I cooperative communication regime. That is, considering a two-way lane scenario, we use the same direction driving vehicles to build clusters, reverse driving vehicles to carry prefetched data, and build clusters to forward prefetched data to improve the data download volume of target vehicles in nonhot scenarios such as highways with the sparse deployment of roadside units, to meet the data volume download demand of in-vehicle users for large files and give guidance for efficient distribution of large file content in highway scenarios.

A MANET (Mobile Ad-hoc Network) is a group of mobile network nodes dynamically forming a network without any pre-existing infrastructure. Multi-path routing protocols in MANETs try to discover and use multiple routes between source and destination nodes. Multipath routing is typically used to reduce average delay, increase transmission reliability, provide load balancing among multiple routes, and improve security and overall QoS (Quality of Service). In this paper, the Cluster-Based Routing Protocol (CBRP), which is a single path MANET protocol is enhanced to use multiple paths. The traffic will be distributed among multiple paths to reduce network traffic congestion and decrease delay. An analytical model is used for multipath and single path CBRP routing protocols in MANETs to estimate the end-to-end delay and queue length. The analytical results show that the average delay and average queue length in multipath CBRP are less than the average delay and queue length in single path CBRP.

The problem of dispersion of mobile robots on a graph asks that n robots initially placed arbitrarily on the nodes of an n-node anonymous graph, autonomously move to reach a final configuration where each node has at most one robot on it. This problem is of significant interest due to its relationship to other fundamental robot coordination problems, such as exploration, scattering, load balancing, relocation of self-driving electric cars to recharge stations, etc. The robots have unique IDs, typically in the range [ 1 , p o l y ( n ) ] and limited memory, whereas the graph is anonymous, i.e., the nodes do not have identifiers. The objective is to simultaneously minimize two performance metrics: (i) time to achieve dispersion and (ii) memory requirement at each robot. This problem has been relatively well-studied when robots are non-faulty. In this paper, we introduce the notion of Byzantine faults to this problem, i.e., we formalize the problem of dispersion in the presence of up to f Byzantine robots. We then study the problem on a ring while simultaneously optimizing the time complexity of algorithms and the memory requirement per robot. Specifically, we design deterministic algorithms that attempt to match the time lower bound ( Ω ( n ) rounds) and memory lower bound ( Ω ( log ⁡ n ) bits per robot). Our main result is a deterministic algorithm that is both time and memory optimal, i.e., O ( n ) rounds and O ( log ⁡ n ) bits of memory required per robot, subject to certain constraints. We subsequently provide results that require less assumptions but are either only time or memory optimal but not both. We also provide a primitive, utilized often, that takes robots initially gathered at a node of the ring and disperses them in a time and memory optimal manner without additional assumptions required.

Recent trends in data-intensive experiments require extensive computing and storage resources that are now handled using cloud resources. Industry experts and researchers use cloud-based services and resources to get analytics of their data to avoid inter-organizational issues including power overhead on local machines, cost associated with maintaining and running infrastructure, etc. This article provides detailed review of selected metrics for cloud computing according to the requirements of data science and big data that includes (1) load balancing, (2) resource scheduling, (3) resource allocation, (4) resource sharing, and (5) job scheduling. The major contribution of this review is the inclusion of these metrics collectively which is the first attempt towards evaluating the latest systems in the context of data science. The detailed analysis shows that cloud computing needs research in its association with data-intensive experiments with emphasis on the resource scheduling area.

The state-of-the-art datacenter load balancing designs commonly optimize bisection bandwidth with homogeneous switching granularity. Their performances surprisingly degrade under mixed traffic containing both short and long flows. Specifically, the short flows suffer from long-tailed delay, while the throughputs of long flows also degrade dramatically due to low link utilization and packet reordering. To solve these problems, we design a traffic-aware load balancing (TLB) scheme to adaptively adjust the switching granularity of long flows according to the load strength of short ones. Under the heavy load of short flows, the long flows use large switching granularity to help short ones obtain more opportunities in choosing short queues to complete quickly. On the contrary, the long flows reroute flexibly with small switching granularity to achieve high throughput. Furthermore, under extremely bursty scenario, we utilize the packet slicing scheme for long flows to release bandwidth for short ones. The experimental results of NS2 simulation and testbed implementation show that TLB significantly reduces the average flow completion time of short flows by 16%-67% over the state-of-the-art load balancers and achieves the high throughput for long flows. Moreover, for extreme bursty case, at the acceptable throughput degradation of long flows, TLB with packet slicing reduces the deadline missing ratio of bursty short flows by up to 80%.

Information-Centric Networking (ICN) provides scalable and efficient content distribution at the Internet scale due to in-network caching and native multicast. To support these features, a content router needs high performance at its data plane, which consists of three forwarding steps: checking the Content Store (CS), then the Pending Interest Table (PIT), and finally the Forwarding Information Base (FIB). In this work, we build an analytical model of the router and identify that CS is the actual bottleneck. Then, we propose a novel mechanism called “NB-Cache” to address CS’s performance issue from a network-wide point of view. In NB-Cache, when packets arrive at a router whose CS is fully loaded, instead of being blocked and waiting for the CS, these packets are forwarded to the next-hop router, whose CS may not be fully loaded. This approach essentially utilizes Content Stores of all the routers along the forwarding path in parallel rather than checking each CS sequentially. NB-Cache follows a design pattern of on-demand load balancing and can be formulated into a non-trivial N-queue bypass model. We use the Markov chain to establish its theoretical base and find an algorithm for automated transition rate matrix generation. Experiments show significant improvement of data plane performance: 70% reduction in round-trip time (RTT) and 130% increase in throughput. NB-Cache decouples the fast packet forwarding from the slower content retrieval thus substantially reducing CS’s heavy dependency on fast but expensive memory.

The emergence of Network Function Virtualization (NFV) and Service Function Chaining (SFC) together enable flexible and agile network management and traffic engineering. Due to the sequential execution nature of SFC, the latency would grow linearly with the number of functions. To resolve this issue, function parallelization has recently been proposed to enable independent functions to work simultaneously. Existing solutions, however, assume all the function instances are installed in the same physical machine and, thus, can be parallelized with only a little overhead. Nowadays, most of the networks deploy function instances in distributed servers for load balancing, parallelization across different servers would, in fact, introduce a non-negligible cost of duplicating or merging packets. Hence, in this work, we propose PPC ( Partial Parallel Chaining ), which only parallelizes functions if parallelization can indeed reduce the latency after considering function placement and the required additional parallelization cost. To this end, we design two schemes, partial parallelism enumeration and instance assignment to identify the optimal partial parallelism that minimizes the latency. Our simulation results show that PPC effectively adapts the degree of parallelism and, hence, outperforms both sequential chaining and full parallelism in any general scenario. Overall, the latency reduction can be up to 47.2% and 35.2%, respectively, as compared to sequential chaining and full parallelism.

In a cloud computing environment, effective scheduling policies and load balancing have always been the aim. An efficient task scheduler must be proficient in a dynamically distributed environment and to the policy of efficient scheduling of jobs based upon the workload. In this research, a novel hybrid heuristic algorithm is developed for balancing the load among cloud nodes. This is achieved by hybridizing the existing ant colony optimization (ACO), artificial bee colony algorithm (ABC), and AHP (analytical hierarchy process) algorithm. The AHP algorithm and the artificial bee colony (ABC) algorithm is used for figuring out the best servers suitable for a particular job, and the ant colony algorithm is used to find the most efficient path to that particular server. The proposed algorithm is better in resource utilization. It also performs better load balancing, which keeps on improving with time. The result analysis shows better average response time and better average makespan time compared to other two existing algorithms.

Recently, with the rapid development of the fifth-generation (5G) communication technology, various new services such as virtual reality (VR), augmented reality (AR), and video conferencing, etc. have emerged. To achieve optimized quality of service (QoS), high load and low latency, we propose an effective mobile edge caching strategy. Specifically: first, a cache prefetching algorithm is proposed to improve cache hit rate. second, the load balancing algorithm based on the maximum distribution weighted entropy is designed to improve bandwidth utilization and avoid mobile network congestion; third, the optimized cache replacement policy is built based on the content future popularity to reduce access latency and backhaul link pressure. With these proposed cache prefetching and replacement algorithm, the access latency can be reduced significantly, the load balance can be maintained and cache hit ratio can be improved obviously in 5G campus networks. This mobile edge caching policy is an effective improvement of the existing strategy, offering a strong support for the full realization of the 5G network potential.

Cloud service providers use load balancing algorithms in order to avoid Service Level Agreement Violations (SLAVs) and wasted energy consumption due to host over- and under-utilization, respectively. Load balancing algorithms migrate VMs between hosts in order to balance host loads. Any Virtual Machines (VMs) that are migrated experience performance degradation which results in lower Quality of Service (QoS) and can possibly result in SLAVs. Hence, an optimal load balancing method should reduce the number of over- and under-utilized hosts with a minimal number of VM migrations. One of the metrics used previously in the literature for evaluating load balancing stated that it equally considered SLAVs caused by both over-utilized hosts and migrations. However, in this paper, we show that, in fact, this metric values keeping the number of migrations low at the expense of an increased number of over-utilized hosts. This disparity is demonstrated by simulation of Google, PlanetLab and Azure data sets in CloudSim. This metric may suit public cloud providers which are focused on minimizing SLAVs and keeping energy costs low, but does not consider the QoS of customer VMs. We propose an alternative metric that considers QoS for the VMs. This alternative metric considers not only performance loss during migration, but also performance degradation due to host over-utilization. Private cloud providers, e.g., IT services within large organizations, often value the performance of their “customer” VMs, i.e., the QoS their organization receives, as well as traditional cloud provider costs, i.e., energy and SLAV costs. Hence, our alternative metric would be more appropriate in these scenarios. We compare and contrast load balancing methods using both the existing, biased metric and our new alternative metric.

Space-Air-Ground networks, as a successful solution for global communications, have drawn extensive attention. Since satellite as the space communications are influenced by geographical features, traffic congestion may occur in densely populated cities, leading to serious latency and throughput collapses. To solve this problem, we first introduce Software Defined Networks (SDN) into satellites and then propose a load balancing approach to improve communication performances. However, the SDN Space-Air-Ground networks need a cooperate communication process, we consider this problem as an SDNenabled routers placement problem, so that we can decide the exactly traditional routers are upgrading to ensure the robustness of hybrid networks. Based on the hybrid Space-Air-Ground networks, we design a load balancing strategy to distribute traffic among networks optimally, so that the total congested links are minimized in our proposed model. For better performance in Space-Air-Ground networks, we propose a resilient congestion estimate scheme. Detouring traffic around the nearby satellites so as to optimize traffic load, is decided based on the network condition. Simulation results illustrate that our proposed method outperforms traditional satellite network methods in the performance of load balancing.