Dynamic Load Balancing Method Research Articles

Proposed Semi-Distributed Framework to Enhance Web Performance

Server load balancing is a technique of circulating client requests across a group of servers. Static and Dynamic load balancing methods are used to distribute the workload equitably over every node of the system. Dynamic load balancing is adaptive in nature and performs load distribution at run time, which makes it more suitable for systems where workload is unpredictable, as compared to static load balancing methods. Two approaches namelyCompletely Distributed and a proposed approach SemiDistributed are tested against two applications – database and ftp. Efficiency of models is tested using various parameters of FTP, IP, and TCP. Simulation results show that semi-distributed method provides better system efficiency and performance.

A low-interference time-domain N-continuous OFDM scheme

In order to achieve dynamic load balancing based on data flow level, in this paper, we apply SDN technology to the cloud data center, and propose a dynamic load balancing method of cloud center based on SDN. The approach of using the SDN technology in the current task scheduling flexibility, accomplish real-time monitoring of the service node flow and load condition by the OpenFlow protocol. When the load of system is imbalanced, the controller can allocate globally network resources. What's more, by using dynamic correction, the load of the system is not obvious tilt in the long run. The results of simulation show that this approach can realize and ensure that the load will not tilt over a long period of time, and improve the system throughput.

An efficient approach to collaborative simulation of variable structure systems on multi-core machines

Complex variable-structure systems (CVSSs) are a common type of complex systems that exhibit changes both at structural and behavior levels. Simulations of CVSSs challenge current collaborative execution methods with increasingly big and complex models. The emergence of multi-core paradigm presents an exciting opportunity to address such challenge, so an advanced parallel simulator under multi-core environments is proposed. The simulator: (1) provides thread simulation kernels and five kinds of management services to support dynamic model structure flexibly; (2) can explore both inherent and dynamic parallelism among models based on interaction relations, and employ the multi-thread paradigm to gain good speedup; (3) adopts an efficient dynamic load-balancing method, which can migrate models among cores with very low cost and support dynamic core allocation on demand, to address evident load-imbalance problems brought by variable-structure. The experiments show that structure changes can be supported while up to 23 % performance increase can be gained.

Dynamic Load Balancing using Graphics Processors

To get maximum performance on the many- core graphics processors, it is important to have an even balance of the workload so that all processing units contribute equally to the task at hand. This can be hard to achieve when the cost of a task is not known beforehand and when new sub-tasks are created dynamically during execution. Both the dynamic load balancing methods using Static task assignment and work stealing using deques are compared to see which one is more suited to the highly parallel world of graphics processors. They have been evaluated on the task of simulating a computer move against the human move, in the famous four in a row game. The experiments showed that synchronization can be very expensive, and those new methods which use graphics processor features wisely might be required.

Implementation and Evaluation of an AMR Framework for FDM Applications

In order to e xecute various finite-difference method applications on large-scale parallel computers with a reasonable cost of computer resources, a framework using an adaptive mesh refinement (AMR) technique has been developed. AMR can realize high -resolution simulations while saving computer resources by generating and removing hierarchica l grids dynamically. In the AMR frame work, a dynamic domain decomposition (DDD) technique, as a dynamic load balancing method, is also implemented to correct the computational load imbalance between each process associated with parallelization. By performing a 3D AM R test simulation, it is confirmed that dynamic load balancing can be achieved and execution time can be reduced by introducing the DDD technique.

MDSLB: A new static load balancing method for parallel molecular dynamics simulations

Large-scale parallelization of molecular dynamics simulations is facing challenges which seriously affect the simulation efficiency, among which the load imbalance problem is the most critical. In this paper, we propose, a new molecular dynamics static load balancing method (MDSLB). By analyzing the characteristics of the short-range force of molecular dynamics programs running in parallel, we divide the short-range force into three kinds of force models, and then package the computations of each force model into many tiny computational units called “cell loads”, which provide the basic data structures for our load balancing method. In MDSLB, the spatial region is separated into sub-regions called “local domains”, and the cell loads of each local domain are allocated to every processor in turn. Compared with the dynamic load balancing method, MDSLB can guarantee load balance by executing the algorithm only once at program startup without migrating the loads dynamically. We implement MDSLB in OpenFOAM software and test it on TianHe-1A supercomputer with 16 to 512 processors. Experimental results show that MDSLB can save 34%–64% time for the load imbalanced cases.

A new step toward load balancing based on competency rank and transitional phases in Grid networks

Grid computing is a network of software–hardware capabilities. It serves as a comprehensive and complete system for organizations by which the maximum utilization from resources is achieved. Resource distribution in a heterogeneous and unstable environment and also effective load distribution among these resources are the important and difficult problems in Grid networks. Using dynamic and static algorithms or searching tree and Branch and Bound algorithm are considered to be among the available methods to reach the load balancing in Grid networks. This paper presents a new method for dynamic load balancing. In this method, we use the subtraction of forward and backward ants as a competency rank to take the priority of the sites, and we use a control word to search the suitable resource as well. Our main purpose is to devote jobs to the existing resources based on their processing power. Simulation results show that the proposed method can reduce the total completion time and also total tardiness to get the load balancing. The cost of using resources as an effective factor in load balancing is also observed.

Periodic hierarchical load balancing for large supercomputers

Large parallel machines with hundreds of thousands of processors are becoming more prevalent. Ensuring good load balance is critical for scaling certain classes of parallel applications on even thousands of processors. Centralized load balancing algorithms suffer from scalability problems, especially on machines with a relatively small amount of memory. Fully distributed load balancing algorithms, on the other hand, tend to take longer to arrive at good solutions. In this paper, we present an automatic dynamic hierarchical load balancing method that overcomes the scalability challenges of centralized schemes and longer running times of traditional distributed schemes. Our solution overcomes these issues by creating multiple levels of load balancing domains which form a tree. This hierarchical method is demonstrated within a measurement-based load balancing framework in Charm++. We discuss techniques to deal with scalability challenges of load balancing at very large scale. We present performance data of the hierarchical load balancing method on up to 16,384 cores of Ranger (at the Texas Advanced Computing Center) and 65,536 cores of Intrepid (the Blue Gene/P at Argonne National Laboratory) for a synthetic benchmark. We also demonstrate the successful deployment of the method in a scientific application, NAMD, with results on Intrepid.

A Compound Parallel Btree for High Scalability and Availability on Chained Declustering Parallel Systems

Scalability and availability are the key features of parallel database systems. To realize scalability, many dynamic load-balancing methods with data placement and parallel index structures on shared-nothing parallel infrastructure have been proposed. Data migration with range-partitioned placement using a parallel Btree is one solution. The combination of range partitioning and chained declustered replicas provides high availability (HA) while preserving scalability. However, independent treatment of the primary and backup data in each node requires long failover times. We propose a novel method for the compound treatment of chained declustered replicas using a parallel Btree, termed the Fat-Btree. In the proposed method, a single Fat-Btree provides access paths to both the primary and backup data of all processor elements (PEs), which greatly reduces failover time. Moreover, these access paths overlap between two neighboring PEs, which enables dynamic load balancing without physical data migration by dynamically redirecting the access paths. In addition, this compound treatment improves memory space utilization to enable index processing with good scalability. Experiments using PostgreSQL on a 160-node PC cluster demonstrate the effectiveness of the high scalability and availability of our proposed method.

Parallel space‐filling curve generation through sorting

AbstractIn this paper we consider the scalability of parallel space‐filling curve generation as implemented through parallel sorting algorithms. Multiple sorting algorithms are studied and results show that space‐filling curves can be generated quickly in parallel on thousands of processors. In addition, performance models are presented that are consistent with measured performance and offer insight into performance on still larger numbers of processors. At large numbers of processors, the scalability of adaptive mesh refined codes depends on the individual components of the adaptive solver. One such component is the dynamic load balancer. In adaptive mesh refined codes, the mesh is constantly changing resulting in load imbalance among the processors requiring a load‐balancing phase. The load balancing may occur often, requiring the load balancer to perform quickly. One common method for dynamic load balancing is to use space‐filling curves. Space‐filling curves, in particular the Hilbert curve, generate good partitions quickly in serial. However, at tens and hundreds of thousands of processors serial generation of space‐filling curves will hinder scalability. In order to avoid this issue we have developed a method that generates space‐filling curves quickly in parallel by reducing the generation to integer sorting. Copyright © 2007 John Wiley & Sons, Ltd.

A refinement-tree based partitioning method for dynamic load balancing with adaptively refined grids

The partitioning of an adaptive grid for distribution over parallel processors is considered in the context of adaptive multilevel methods for solving partial differential equations. A partitioning method based on the refinement-tree is presented. This method applies to most types of grids in two and three dimensions. For triangular and tetrahedral grids, it is guaranteed to produce connected partitions; no other partitioning method makes this guarantee. The method is related to the OCTREE method and space filling curves. Numerical results comparing it with several popular partitioning methods show that it computes partitions in an amount of time similar to fast load balancing methods like recursive coordinate bisection, and with mesh quality similar to slower, more optimal methods like the multilevel diffusive method in ParMETIS.

A Dynamic Load-Balancing Parallel Search for Enumerative Robot Path Planning

We present a parallel formulation for enumerative search in high dimensional spaces and apply it to planning paths for a 6-dof manipulator robot. Participating processors perform local A* search towards the goal configuration. To exploit all the processors at their maximum capacity at all times, a dynamic load-balancing scheme matches idle and busy processors for load transfer. For comparison purposes, we have also implemented an existing parallel static load-balancing formulation based on regular domain decomposition. Both methods achieved almost linear speed-up in our experiments. The two methods follow different search strategies in parallel and the implementation of the existing method (with tuned space decomposition) was more time efficient on average. However, the planning time of that method is highly dependent on the distribution of the search space among the processors and its tuned decomposition varies for different obstacle placements. Empirical selection of the space decomposition parameters for the existing method does not guarantee minimal planning time in all environments and leads to slower planning than our dynamic load-balancing method in some cases. The performance of the developed dynamic method is independent of the obstacle placements and the method can achieve consistent speed-up in all environments.

MMORPG에서의 부하 분산을 위한 가상 영역 정보 기반 동적 지역 분할

MMORPG(Massively Multiplayer Online Role-Playing Game)은 대규모의 인원들이 동일한 가상공간에서 동일한 시각에 상호작용하면서 활동하는 온라인 롤플레잉 게임이다. 대부분의 경우, 서버와 대역폭 등에서 상당한 수준의 하드웨어 사양과 지원 스태프를 필요로 한다. 수많은 개발자들의 노력에도 불구하고 사용자들은 여전히 플레이어의 과밀, 랙, 그리고 부족한 지원 등에 대해 불만을 가지는 경우가 많다. 본 논문에서는 MMORPG에서의 부하 분산을 위한 동적 지역 분할 방법을 제안한다. Field, sector group, sector, cell 등으로 구성되는 가상 영역 정보를 가진 VML(Virtual Map Layer)를 이용하여 지역을 분할함으로써 부하 분산을 시도한다. 실험 결과, 제안된 기법이 <TEX>$23^{\sim}67%$</TEX> 정도의 부하를 줄이는 효과를 가져왔다. 맵 데이터, 자원의 상태, 그리고 사용자들의 행동 패턴의 변화 등에 대해서는 VML에 대한 간단한 수정으로 유연하게 적응할 수 있다. A MMORPG(Massively Multiplayer Online Role-Playing Game) is an online role-playing game in which a large number of players can interact with each other in the same world at the same time. Most of them require significant hardware requirements(e.g., servers and bandwidth), and dedicated support staff. Despite the efforts of developers, users often cite overpopulation, lag, and poor support as problems of games. In this paper, a dynamic load balancing method for MMORPGS is proposed. It tries to adapt to dynamic change of population by using dynamic map-partition method with VML(Virtual Map Layer) which consists of fields, sector groups, sectors, and cells. From the experimental results, our approach achieves about <TEX>$23^{\sim}67%$</TEX> lower loads for each field server. By the modification to Virtual Area Layer, we can easily manage problems that come from changes of map data, resources' status, and users' behavior pattern.

A parallel cost‐based abductive reasoning system on workstation cluster

AbstractThis paper proposes a method for dynamic load balancing which works efficiently in the workstation cluster environment. A parallel abductive reasoning system based on the proposed method is also proposed. In the workstation cluster environment, parallel processing is performed using workstations with various processing abilities and loads. Thus, dynamic load balancing must consider the performance and loading conditions of the workstations. This paper proposes a method of estimating the processing ability of the workstation during operation, and also a dynamic load balancing algorithm to balance the estimated conditions. Using the proposed method, a parallel abductive reasoning system which can derive the most preferable explanation for a given observation is implemented. The results of experiments on the system are also reported. © 2006 Wiley Periodicals, Inc. Syst Comp Jpn, 37(3): 80–89, 2006; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/scj.10190

Dynamic load balancing of plasma particle-in-cell simulations: The taskfarm alternative

Particle-in-cell (PIC) simulations frequently exhibit some degree of inhomogeneity in the distribution of particles. When running such codes in parallel, this inhomogeneity results in an imbalance of the particle-associated workload. As a remedy, plasma PIC simulations commonly make use of a dynamic domain decomposition that adjusts itself according to the momentary distribution of particles. We present and evaluate an alternative dynamic load-balancing method: a taskfarm-based scheme, where the total particle-associated work is traded among all processors according to their individual workload. It will be shown to be very efficient for simulations with highly inhomogeneous particle distributions.

A dynamic load balancing method based on network delay for large distributed systems

A dynamic load balancing method based on network delay for large distributed systems

A dynamic load balancing method based on network delay for large distributed systems

The authors propose two dynamic load balancing methods which give consideration to network delays in a large-scale, wide-area distributed system. In general in load balancing, the number of candidate computers for transfers increases as the number of computers being used increases, and so efforts are made to improve response time. There is, however, a concern that performance will drop due to delay overhead when candidate destinations for transfers include computers with large delays. Therefore, in this research the authors configure a collection of computers with low network delays as a cluster. They then propose a clustered dynamic load balancing method that emphasizes load information within a cluster and delay information when transfers are performed between clusters, and a threshold dynamic load balancing method which restricts process destinations by creating a threshold value with respect to the ratio between the delay required for a process transfer and the average processing request time for a process. The results of a performance evaluation performed using simulations show that each proposed method can greatly improve the average response time compared to methods that do not consider delays. In addition, each method has an environment in which it is thought to be particularly effective given its operational characteristics. © 2001 Scripta Technica, Electron Comm Jpn Pt 1, 84(6): 11–21, 2001

Parallel load balancing for dynamic execution environments

The development and implementation of a new method for parallel dynamic load balancing is presented. The load balancer uses an octree-based method to calculate the amount of imbalance and determine a new partitioning for the hybrid mesh. The balancer is designed to redistribute the work when the loads on the processors change due to local mesh adaptation or to changes in the parallel execution system. For the case of a dynamic parallel system, the loads on the processors are determined by run time measurements and the balancer redistributes the work based on these timings. Since the load balancer uses the same octree-based technique employed by the partitioner, the overall change in the partitions is minimized and a reduced amount of data migration is realized. The qualities of the partitions are maintained even after multiple load balancings. The effectiveness of the dynamic load balancer is demonstrated via parallel execution times for adaptive turbulent flow simulations.

RANDOM SEEKING: A GENERAL, EFFICIENT, AND INFORMED RANDOMIZED SCHEME FOR DYNAMIC LOAD BALANCING

We propose a completely general, informed randomized dynamic load balancing method called random seeking (RS) suitable for parallel algorithms with characteristics found in many search algorithms used in artificial intelligence and operations research and many divide-and-conquer algorithms. In it, source processors randomly seek out sink processors for load balancing by flinging "probe" messages. These probes not only locate sinks, but also collect load distribution information which is used to efficiently regulate load balancing activities. We empirically compare RS with a well-known randomized dynamic load balancing method, the random communication (RC) strategy, by using them in parallel best-first branch-and-bound algorithms on up to 512 processors of a parallel system. We find that the RC execution times are more than those of RS by 16–67% when used to perform combined dynamic quantitative and qualitative load balancing, and by 9–74% when used to perform only dynamic quantitative load balancing.

Synchronous Load Balancing in Hypercube Multicomputers with Faulty Nodes

This paper presents a new dynamic load-balancing algorithm for hypercube multicomputers with faulty nodes. The emphasis in our method is on obtaining global load information and performing task migration using “short paths” in a synchronous manner so that a minimal amount of communication overhead is required. To accomplish this, we present an algorithm for constructing a new logical topology from a hypercube topology with faulty nodes. This new topology is used to obtain the global load information and to perform task migration. Simulation results are used to evaluate the performance of our dynamic load balancing method when compared with previous methods