Search

Mobile edge computing (MEC) is one of the efficient technologies to provide satisfying quality of experience (QoE) for emerging computation-intensive applications in internet of things (IoT). However, some new challenges will be encountered when MEC is applied in a large-scale IoT with massive access devices or heavy traffic loads such as load balancing and traffic offloading. Aiming at the solution of these problems, this paper proposes a learning-based traffic control architecture for IoT with MEC. Moreover, a deep-learning-based load balancing framework is developed to control user association in IoT. The user association is determined at each IoT access points (IAP) by the deep neural network (DNN), which is the duplication of the well-trained DNN with the global network information. In addition, we propose a reinforcement-learning-based partial traffic offloading scheme to reduce the traffic origination. The IoT devices are able to independently decide its offloading radio according to the channel quality information, service requirement, and workload of the IAP. Simulation results indicate that the proposed deep-learning-based load balancing scheme is able to achieve uniform traffic distribution, and meanwhile our partial offloading policy can significantly reduce the network traffic. [ABSTRACT FROM AUTHOR]

In cloud computing, a lot of challenges like the server failures, loss of confidentiality, improper workloads, etc. are still bounding the efficiency of cloud systems in real-world scenarios. For this reason, many research works are being performed to overcome the shortcoming of existing systems. Among them, load balancing seems to be the most critical issue that worsen the performance of the cloud sector, and hence there necessitates the optimal load balancing with optimal task scheduling. With the intention of accomplishing optimal load balancing by effectual task deployment, this paper plans to develop an advanced load balancing model with the assistance acquired from the metaheuristic algorithms. Usually, handling of tasks in cloud system is an NP-hard problem and moreover, nonpreemptive independent tasks are crucial in cloud computing. This paper goes with the introduction of a new optimal load balancing model by considering three major objectives: minimum makespan, priority, and load balancing, respectively. Moreover, a new single-objective function is also defined that incorporates all the three objectives mentioned above. Furthermore, the deployment of tasks must be optimal and for this a new hybrid optimization algorithm referred as Firefly Movement insisted WOA (FM-WOA) is introduced. This FM-WOA is the conceptual amalgamation of standard Whale Optimization Algorithm (WOA) and Firefly (FF) algorithm. Finally, the performances of the proposed FM-WOA model is compared over the conventional models with the intention of proving its efficiency in terms of makespan, task completion (priority), and degree of imbalance as well. [ABSTRACT FROM AUTHOR]

Cloud computing brings incipient transmutations in different fields of life and consists of different characteristics and virtualization is one of them. Virtual machine (VM) is one of the main elements of virtualization. VM is a process in which physical server changes into the virtual machine and works as a physical server. When a user sends data or request for data in cloud data center, a situation can occur that may cause the virtual machines to underload data or overload data. The aforementioned situation can lead to failure of the system or delay the user task. Therefore, appropriate load balancing techniques are required to surmount the above two mentioned problems. Load balancing is a technique utilized in cloud computing for management of the resource by a condition such that a maximum throughput is achieved with slightest reaction time and additionally dividing the traffic between different servers or VM so that it can get data without any delay. For the amelioration of load balancing technique in this study, a novel technique is used which is coalescence of BAT and ABC algorithms both of which are nature-inspired algorithms. When the ABC algorithm local search section changes with BAT algorithm local search section, a second modification takes place in the fitness function of BAT algorithm. The proposed technique is known as HBATAABC algorithm. The novel technique implemented by utilizing transfer strategy policy in VM improves the performance of data allocation system of VM in the cloud data center. To check the performance of the proposed algorithm, three main parameters are used which are network average time, network stability and throughput. The performance of the proposed novel technique is verified and tested with the help of cloudsim simulator. The result shows that the suggested modified algorithm increases performance by 1.30% of network average time, network stability and throughput as compared with BAT algorithm, ABC algorithm and RRA algorithm. Nevertheless, the proposed algorithm is more precise and expeditious as compared with the three models. [ABSTRACT FROM AUTHOR]

This paper investigates semi-online hierarchical scheduling problems on two identical machines, with the purpose of minimizing the l p -norm of the machines' loads. We consider two semi-online versions with knowing the total processing time T of all jobs, or knowing the total processing time T i of the jobs of hierarchy i for i = 1 , 2 in advance. For the two semi-online versions, the best possible online algorithms are designed with competitive ratios of φ ∗ and 2 p + 4 p 2 (3) p 1 p , respectively, where φ ∗ is the maximum value of the function φ (x) = (x + 1 2 (1 − x)) p + (1 2 (1 − x)) p x p + (1 − x) p 1 p in x ∈ [ 1 2 , 1 ]. When p = ∞ , our results cover the known results for minimizing the makespan. [ABSTRACT FROM AUTHOR]

Revealing how the interdependency between different networks affects the load-induced cascading failures is a new interesting topic recently. A new cascading failure model for interconnected network system is introduced in this paper, in which the loads of the failed nodes can be transmitted between different layer networks or among each layer network, the ratio is adjusted with a defined parameter T f which can be viewed as a measure of coupling strength. Based on the proposed model, we find that under different attacking strategies and different value of T f , different coupling types have different effects on the robustness of interconnected network systems. We mainly focus on the robustness of interconnected network with random coupling under intentional attack, we find there being a threshold T f ∗ , when T f < T f ∗ , the interconnected network systems will be more vulnerable with larger value of T f ; however, when T f > T f ∗ , if the capacity redundancy of interconnected network is low, increasing T f tends to enhance the robustness of interconnected network system. In addition, we also get the conclusion that increasing the total amount of loads in one layer network or the unbalance of load distribution between different layer networks will decline the robustness of the interconnected network system. We hope our work will shed some light on designing high-robust interconnected network systems or improving the robustness of interconnected network systems. [ABSTRACT FROM AUTHOR]

Load balancing, which redistributes dynamic workloads across computing nodes within cloud to improve resource utilization, is one of the main challenges in cloud computing system. Most existing rule-based load balancing algorithms failed to effectively fuse load data of multi-class system resources. The strategies they used for balancing loads were far from optimum since these methods were essentially performed in a combined way according to load state. In this work, a fuzzy clustering method with feature weight preferences is presented to overcome the load balancing problem for multi-class system resources and it can achieve an optimal balancing solution by load data fusion. Feature weight preferences are put forward to establish the relationship between prior knowledge of specific cloud scenario and load balancing procedure. Extensive experiments demonstrate that the proposed method can effectively balance loads consisting of multi-class system resources. [ABSTRACT FROM AUTHOR]

With the advancement in the science and technology, cloud computing has become a recent trend in environment with immense requirement of infrastructure and resources. Load balancing of cloud computing environments is an important matter of concern. The migration of the overloaded virtual machines (VMs) to the underloaded VM with optimized resource utilization is the effective way of the load balancing. In this paper, a new VM migration algorithm for the load balancing in the cloud is proposed. The migration algorithm proposed (EGSA-VMM) is based on exponential gravitational search algorithm which is the integration of gravitational search algorithm and exponential weighted moving average theory. In our approach, the migration is done based on the migration cost and QoS. The experimentation of proposed EGSA-based VM migration algorithm is compared with ACO and GSA. The simulation of experiments shows that the proposed EGSA-VMM algorithm achieves load balancing and reasonable resource utilization, which outperforms existing migration strategies in terms of number of VM migrations and number of SLA violations. [ABSTRACT FROM AUTHOR]

The scalability of the cloud infrastructure is essential to perform large-scale data processing using MapReduce programming model by automatically provisioning and de-provisioning the resources on demand. The existing MapReduce model shows performance degradation while getting adapted to heterogeneous environments since sufficient techniques are not available to scale the resources on demand and the scheduling algorithms would not cooperate as the resources are configured dynamically. An Auto-Scaling Framework (ASF) is presented in this article to configure the resources automatically based on the current system load in a heterogeneous Hadoop environment. The scheduling of data and task is done in a data-local manner that adapts while new resources are configured, or the existing resources are removed. A monitoring module is integrated with the JobTracker to observe the status of physical machines, compute the system load and provide automated provisioning of the resources. Then, Replica Tracker is utilized to track the replica objects for efficient scheduling of the task in the physical machines. The experiments are conducted in a commercial cloud environment using diverse workload characteristics, and the observations show that the proposed framework outperforms the existing scheduling mechanisms by the performance metrics such as average completion time, scheduling time, data locality, resource utilization and throughput. [ABSTRACT FROM AUTHOR]

The routing strategy plays a very important role in complex networks such as Internet system and Peer-to-Peer networks. However, most of the previous work concentrates only on the path selection, e.g. Flooding and Random Walk, or finding the shortest path (SP) and rarely considering the local load information such as SP and Distance Vector Routing. Flow-based Routing mainly considers load balance and still cannot achieve best optimization. Thus, in this paper, we propose a novel dynamic routing strategy on complex network by incorporating the local load information into SP algorithm to enhance the traffic flow routing optimization. It was found that the flow in a network is greatly affected by the waiting time of the network, so we should not consider only choosing optimized path for package transformation but also consider node congestion. As a result, the packages should be transmitted with a global optimized path with smaller congestion and relatively short distance. Analysis work and simulation experiments show that the proposed algorithm can largely enhance the network flow with the maximum throughput within an acceptable calculating time. The detailed analysis of the algorithm will also be provided for explaining the efficiency. [ABSTRACT FROM AUTHOR]

This paper is aimed at constructing robust multilayer networks against cascading failure. Considering link protection strategies in reality, we design a cascading failure model based on load distribution and extend it to multilayer. We use the cascading failure model to deduce the scale of the largest connected component after cascading failure, from which we can find that the performance of four kinds of load distribution strategies associates with the load ratio of the current edge to its adjacent edge. Coupling preference is a typical characteristic in multilayer networks which corresponds to the network robustness. The coupling preference of multilayer networks is divided into two forms: the coupling preference in layers and the coupling preference between layers. To analyze the relationship between the coupling preference and the multilayer network robustness, we design a construction algorithm to generate multilayer networks with different coupling preferences. Simulation results show that the load distribution based on the node betweenness performs the best. When the coupling coefficient in layers is zero, the scale-free network is the most robust. In the random network, the assortative coupling in layers is more robust than the disassortative coupling. For the coupling preference between layers, the assortative coupling between layers is more robust than the disassortative coupling both in the scale free network and the random network. [ABSTRACT FROM AUTHOR]

Recently the robustness of coupled network under cascading failure has attracted a lot of attention. In this paper, we investigate the cascading failure of the interconnected weighted networks based on the state of link. The load on one link is defined by a function of the strength of the two nodes at the ends of that link, using four intentional attack strategies, we study the invulnerability of the interconnected weighted networks when cascading failure occurs. Our studies show that when the link with highest load is attacked, the damage to the network will be more serious by attacking the inner-link with highest load than that caused by attacking the coupling link with highest load, and no matter how the coupling links distribute, there are two thresholds. In addition, we find that the larger the weight increment in the model or the smaller the network's mean clustering coefficient, the stronger the ability of the network to resist cascading failure when the inner-link with highest load is attacked, while the weaker the ability of the network to suppress the cascading failure when the inner-link with lowest load is attacked. [ABSTRACT FROM AUTHOR]

Cascading failures in networked systems often lead to catastrophic consequence. Defending cascading failure propagation by employing local load redistribution method is an efficient way. Given initial load of every node, the key of improving network robustness against cascading failures is to maximally defend cascade propagation with minimum total extra capacity of all nodes. With finite total extra capacity of all nodes, we first discuss three general extra capacity distributions including degree-based distribution (DD), average distribution (AD) and random distribution (RD). To sufficiently use the total spare capacity (SC) of all neighboring nodes of a failed node, then we propose a novel SC-based local load redistribution mechanism to improve the cascade defense ability of network. We investigate the network robustness against cascading failures induced by a single node failure under the three extra capacity distributions in both scale-free networks and random networks. Compared with the degree-based (DB) local load redistribution method, our SC method achieves higher robustness under all of the three extra capacity distributions. The extensive simulation results can well confirm the effectiveness of the SC local load redistribution method. [ABSTRACT FROM AUTHOR]

We study the problem of universal resilience patterns in complex networks against cascading failures. We revise the classical betweenness method and overcome its limitation of quantifying the load in cascading model. Considering that the generated load by all nodes should be equal to the transported one by all edges in the whole network, we propose a new method to quantify the load on an edge and construct a simple cascading model. By attacking the edge with the highest load, we show that, if the flow between two nodes is transported along the shortest paths between them, then the resilience of some networks against cascading failures inversely decreases with the enhancement of the capacity of every edge, i.e. the more capacity is not always better. We also observe the abnormal fluctuation of the additional load that exceeds the capacity of each edge. By a simple graph, we analyze the propagation of cascading failures step by step, and give a reasonable explanation of the abnormal fluctuation of cascading dynamics. [ABSTRACT FROM AUTHOR]

In the modern society, most networked critical infrastructures are coupled together and can be modeled as interconnected networks. Cascading failures caused by overload in interconnected networks are extremely complicated and still lack of corresponding researches. It is crucial to explore a systematic design and load management method for complex interconnected network. In this paper, a probabilistic heterogeneous capacity allocation model and a preventive load redistribution model are firstly proposed. Based on the novel models failure cascades in interconnected networks are studied in detail. Numerical simulations and analysis results show that robustness of interconnected network depends on the network topology, capacity allocation pattern and load management strategy. The novel models have guiding significance for the design and optimal management of robust interconnected infrastructure networks. [ABSTRACT FROM AUTHOR]

A routing aggregation (RA) is proposed for load balancing network-on-chip (NoC). The computing nodes with dense traffic and long distance in network are gathered into the same routing node to form a super router. A load balancing routing algorithm for super router is presented to improve the overall performance of NoC. A simulation platform using System C is presented to confirm the feasibility of the proposed design in 2D mesh. The simulation results show that the proposed RA design can reduce the average packet latency and the standard deviation of host link utilization 8% and 33%, respectively compared with the reported routing methods. The area cost and power consumption compared with the reported schemes are 22% and 12% less, respectively. [ABSTRACT FROM AUTHOR]

This paper concerns the asymmetric atomic selfish routing game for load balancing in ring networks. In the selfish routing, each player selects a path in the ring network to route one unit traffic between its source and destination nodes, aiming at a minimum maximum link load along its own path. The selfish path selections by individuals ignore the system objective of minimizing the maximum load over all network links. This selfish ring load (SRL) game arises in a wide variety of applications in decentralized network routing, where network performance is often measured by the price of anarchy (PoA), the worst-case ratio between the maximum link loads in an equilibrium routing and an optimal routing. It has been known that the PoA of SRL with respect to classical Nash Equilibrium (NE) cannot be upper bounded by any constant, showing large loss of efficiency at some NE outcome. In an effort to improve the network performance in the SRL game, we generalize the model to so-called SRL with collusion (SRLC) which allows coordination within any coalition of up to k selfish players on the condition that every player of the coalition benefits from the coordination. We prove that, for m-player game on n-node ring, the PoA of SRLC is n - 1 when k ≤ 2, drops to 2 when k = 3 and is at least 1 + 2/m for k ≥ 4. Our study shows that on one hand, the performance of ring networks, in terms of maximum load, benefits significantly from coordination of self-interested players within small-sized coalitions; on the other hand, the equilibrium routing in SRL might not reach global optimum even if any number of players can coordinate. [ABSTRACT FROM AUTHOR]

The wireless sensor networks (WSNs) are formed by a large number of sensor nodes working together to provide a specific duty. However, the low energy capacity assigned to each node prompts users to look at an important design challenge which is lifetime maximization. Especially the availability of nodes, the sensor coverage, and the connectivity have been included in discussions on network lifetime. Therefore, designing effective techniques that conserve scarce energy resources is a critical issue in WSNs. In this regard, we are interested in developing various mechanisms to save energy based on the constraints involved for energy consumption in WSNs. Three mechanisms are proposed to reduce the number of control packets responsible for path discovery, optimize diffusion area, and balance the load distribution in the network. [ABSTRACT FROM AUTHOR]

In 2012, Page presented a sequential combinatorial generation algorithm for generalized types of restricted weak integer compositions called second-order restricted weak integer compositions. Second-order restricted weak integer compositions cover various types of restricted weak integer compositions of n parts such as integer compositions, bounded compositions, and part-wise integer compositions. In this paper, we present a parallel algorithm that derives from our parallelization of Page's sequential algorithm with a focus on load balancing for shared memory machines. [ABSTRACT FROM AUTHOR]

In a large multiprocessor server platform using multicore chips, the scheduler often migrates a thread or process, in order to achieve better load balancing or ensure fairness among competing scheduling entities. Each migration incurs a severe performance impact from the loss of cache and Translation Lookaside Buffer (TLB) footprints and subsequent higher cache misses and page walks. Such impact is likely to be more severe in virtualized environments, where high over-subscription of CPUs is very common for server consolidation workloads or virtual desktop infrastructure deployment, causing frequent migrations and context switches. We demonstrate the performance benefit of preserving a portion of L2 cache-in particular, MRU cache lines-and warming the destination L2 cache by prefetching those cache lines under different migration scenarios. We observed a 1.5-27% reduction in CPI (cycles per instruction) following a migration. We also study the effectiveness of preserving TLB entries over a context switch or migration. [ABSTRACT FROM AUTHOR]