Adaptive Load Balancing Algorithm Research Articles

The construction of a shared resource base of enterprise financial informationization for intelligent terminal platform

Abstract This paper uses an adaptive load-balancing algorithm to calculate an intelligent terminal platform on the basis of an enterprise financial information-sharing model. Firstly, a new server performance evaluation index is established based on the hierarchical analysis method, and the load performance of each node is reacted using the influence factors. Then a judgment matrix is constructed for the set of evaluation factors based on the change in the average load utilization rate of nodes. Finally, the results of the matrix are normalized to derive the evaluation weights of each node of the server, which completes the construction of the intelligent terminal platform. An enterprise’s quarterly bills for 2018-2020 were selected for analysis, and after running the financial information sharing, the processing time of a single voucher could be saved by 0.3 hours. It shows that the constructed information-sharing resource base can not only significantly reduce operation costs but also access the business data of subsidiaries at any time to effectively prevent operation risks.

An Efficient Scheme for Coupling OpenMC and FLUENT with Adaptive Load Balancing

This paper develops a multi-physics interface code MC-FLUENT to couple the Monte Carlo code OpenMC with the commercial computational fluid dynamics code ANSYS FLUENT. The implementations and parallel performances of block Gauss–Seidel-type and block Jacobi-type Picard iterative algorithms have been investigated. In addition, this paper introduces two adaptive load-balancing algorithms into the neutronics and thermal-hydraulics coupled simulation to reduce the time cost of computation. Considering that the different scalability of OpenMC and FLUENT limits the performance of block Gauss–Seidel algorithm, an adaptive load-balancing algorithm that can increase the number of nodes dynamically is proposed to improve its efficiency. Moreover, with the natural parallelism of block Jacobi algorithm, another adaptive load-balancing algorithm is proposed to improve its performance. A 3 x 3 PWR fuel pin model and a 1000 MWt ABR metallic benchmark core were used to compare the performances of the two algorithms and verify the effectiveness of the two adaptive load-balancing algorithms. The results show that the adaptive load-balancing algorithms proposed in this paper can greatly improve the computing efficiency of block Jacobi algorithm and improve the performance of block Gauss–Seidel algorithm when the number of nodes is large. In addition, the adaptive load-balancing algorithms are especially effective when a case demands different computational power of OpenMC and FLUENT.

A Unified Solution for Gateway and In-Network Traffic Load Balancing in Multihop Data Collection Scenarios

Data collection is a fundamental task of wireless sensor networks to support a variety of applications, such as remote monitoring, and emergency response, where collected information is relayed to an infrastructure network via packet gateways for processing and decision making. In large-scale monitoring scenarios, data packets need to be relayed over multihop paths to the gateways, and sensors are often randomly deployed, causing local node density differences. As a result, imbalance in data traffic load on the gateways is likely to occur. Furthermore, due to dynamic network conditions and differences in sensor data generation rates, congestion on some data paths is also often experienced. Numerous studies have focused on the problem of in-network traffic load balancing, whereas a few works have aimed at equalizing the loads on gateways. However, there is a potential tradeoff between these two problems. In this paper, the dual objective of gateway and in-network load balancing is addressed, and the reactive and adaptive load-balancing (RALB) algorithm is presented. RALB is proposed as a generic solution for multihop networks and mesh topologies, particularly in large-scale remote monitoring scenarios, to balance traffic loads.

ALBL: an adaptive load balancing algorithm for distributed web systems

This paper presents an adaptive load balancing algorithm (ALBL) for cluster-based web systems. The balancing policy is based on two criteria: HTTP process time and network delay. The former describes web server ability to process a forthcoming request, while the latter tries to estimate network conditions. Periodic criteria calculations are performed by agents at the web switch and the weight estimation process is transparent to web servers enhancing therefore distributed system's scalability. We put to test our implementation against known blind selection balancing algorithms used at web-farms such as: Round Robin, stateful ones such as Least Connections and adaptive ones such as Least Load. We also put to test performance and scalability of previous algorithms. From our testbed scenario results we show that ALBL algorithm outperforms stateless and stateful algorithms and also presents significant performance gains towards adaptive algorithms. We also show that our algorithm scales well as the number of balancing servers and web cluster's requests rate increases. We also pinpoint ALBL algorithm ability to predict network conditions and web servers load without the use of feedback information obtained by web servers but frequent service and network probes towards web servers issued by the web switch.

An Adaptive Load Balancing Method of Adjustment Based on Cloud Computing Resources

By combining LVS cluster architecture and could computing technology, system architecture of the cloud computing service platform is proposed. Cloud computing technology is suitable to support large-scale applications with flash crowds by support elastic amounts of bandwidth and storage resource etc. But traditional algorithms of load balancing provided by LVS are unsuitable for the proposed service platform, because these algorithms are designed for static server resource provided by traditional cluster technology. Taking both the overall utilization rate of server resources and the active connections of the server into counter, an adaptive adjustable load balancing algorithms (Least Comprehensive Utilization and Connection Scheduling algorithm, called LUCU) is proposed in this paper. According the utilization of cloud resource and the users demand, automatic switching between Round Robin (RR) algorithm and LUCU algorithm is achieved. When the cloud capacities are not able to meet the instantaneous demands, LUCU is chosen instead of RR. The proposed platform and algorithm are verified and evaluated using large-scare simulation experiments. The test results show that the equilibrium load is nearly achieved by adopting the proposed algorithms.

Novel Implementation of Recursive Discrete Wavelet Transform for Real Time Computation with Multicore Systems on Chip (SOC)

The discrete wavelet Transform (DWT) has been studied and developed in various scientific and engineering fields. Its multi-resolution and locality nature facilitates application required for progressiveness in capturing high-frequency details. However, when dealing with enormous data volume, the performance may drastically reduce. The multi-resolution sub-band encoding provided by DWT enables for higher compression ratios, and progressive transformation of signals. The widespread usage of the DWT has motivated the development of fast DWT algorithms and their tuning on all sorts of computer systems. However, this transformation comes at the expense of additional computational complexity. Achieving real-time or interactive compression/de-compression speed, therefore, requires a fast implementation of DWT that leverages emerging parallel hardware systems. The recent advancement in the consumer level multicore hardware is equipped with Single Instruction and Multiple Data (SIMD) power.In this study, Parallel Discrete Wavelet Transform has been developed with novel Adaptive Load Balancing Algorithm (ALBA). The DWT is parallelized, partitioned, mapped and scheduled on single core and Multicore. The Parallel DWT is developed in C# for single and Intel Quad cores as well as the combination of C and CUDA is implemented on GPU. This brings the significant performance on a consumer level PC without extra cost.

Adaptive load balancing algorithm for multiple homing mobile nodes

In places where mobile users can access multiple wireless networks simultaneously, a multipath scheduling algorithm can benefit the performance of wireless networks and improve the experience of mobile users. However, existing literature shows that it may not be the case, especially for TCP flows. According to early investigations, there are mainly two reasons that result in bad performance of TCP flows in wireless networks. One is the occurrence of out-of-order packets due to different delays in multiple paths. The other is the packet loss which is resulted from the limited bandwidth of wireless networks. To better exploit multipath scheduling for TCP flows, this paper presents a new scheduling algorithm named Adaptive Load Balancing Algorithm (ALBAM) to split traffic across multiple wireless links within the ISP infrastructure. Targeting at solving the two adverse impacts on TCP flows, ALBAM develops two techniques. Firstly, ALBAM takes advantage of the bursty nature of TCP flows and performs scheduling at the flowlet granularity where the packet interval is large enough to compensate for the different path delays. Secondly, ALBAM develops a Packet Number Estimation Algorithm (PNEA) to predict the buffer usage in each path. With PNEA, ALBAM can prevent buffer overflow and schedule the TCP flow to a less congested path before it suffers packet loss. Simulations show that ALBAM can provide better performance to TCP connections than its other counterparts.

A Novel Globally Adaptive Load-Balanced Routing Algorithm for Torus Interconnection Networks

A globally adaptive load-balanced routing algorithm for torus interconnection networks is proposed. Unlike previously published algorithms, this algorithm employs a new scheme based on collision detection to handle deadlock, and has higher routing adaptability than previous algorithms. Simulation results show that our algorithm outperforms previous algorithms by 16% on benign traffic patterns, and by 10% to 21% on adversarial traffic patterns.

Geographic forwarding and adaptive load balancing in wireless sensor networks

In the future sensor nodes will either have GPS capabilities, or be able to run localization protocols in order to get an accurate estimation of their position. These features would make geographic convergecasting a feasible and promising solution for WSNs. Key for the actual implementation of WSNs is the design of cross-layer solutions which consider explicitly the resource constraints of sensor nodes, especially in terms of available energy, limited memory and processing capabilities. We present a new protocol, named ALBA ( Adaptive-Load Balanced Algorithm) , which uses a cross-layer approach, pursuing the following goals: i ) ALBA minimizes the number of hops required to reach the sink through a geographic forwarding scheme; ii ) traffic is distributed evenly in the network, favoring nodes that are experiencing low congestion, while avoiding overloaded regions; iii ) channel access efficiency is optimized through an adaptive back-to-back transmission of "bursts of packets" once a relay is selected, and iv ) nodes are progressively made aware of routes to redirect the packets to the sink even when a connectivity hole occurs, i.e., when some nodes are not able to deliver packets to the sink because of lack of next-hop relay nodes closer to the sink.

Adaptive Load Balancing of Parallel Applications with Multi‐Agent Reinforcement Learning on Heterogeneous Systems

We report on the improvements that can be achieved by applying machine learning techniques, in particular reinforcement learning, for the dynamic load balancing of parallel applications. The applications being considered in this paper are coarse grain data intensive applications. Such applications put high pressure on the interconnect of the hardware. Synchronization and load balancing in complex, heterogeneous networks need fast, flexible, adaptive load balancing algorithms. Viewing a parallel application as a one‐state coordination game in the framework of multi‐agent reinforcement learning, and by using a recently introduced multi‐agent exploration technique, we are able to improve upon the classic job farming approach. The improvements are achieved with limited computation and communication overhead.

Performance modeling and analysis of load balancing policies with priority queueing

In this article, we study an adaptive load-balancing algorithm in the homogeneous distributed systems in which only local status information is used. The parameters affecting the performance of the load-balancing algorithm are investigated. To analyze the effects of service disciplines on load balancing, we study two classes of service disciplines, impartial discipline and partial discipline. In impartial discipline, all tasks in the system are treated alike. Partial disciplines divide tasks into two classes, local tasks and remote tasks, and then assign different priorities to them. Five partial disciplines with different priority assignment rules are compared. The numerical results are presented and used to shed light on the characteristics of the load-balancing process.