A partially asynchronous and iterative algorithm for distributed load balancing

This paper deals with the numerical solution of financial applications, more specifically the computation of American and European options derivatives modeled by boundary value problems. In such applications we have to solve large-scale algebraic linear systems. We concentrate on synchronous and asynchronous parallel iterative algorithms carried out on Grid’5000, by using an experimental peer-to-peer platform. The properties of the operators arising in the discretized problem ensure the convergence of the parallel iterative synchronous and asynchronous algorithms. Different detection convergence algorithms are tested for asynchronous iterations. Computational experiments performed on distributed architectures are presented and analyzed.

Distributed file systems are key building blocks for cloud computing applications based on the MapReduce programming paradigm. In such file systems, nodes simultaneously serve computing and storage functions; a file is partitioned into a number of chunks allocated in distinct nodes so that MapReduce tasks can be performed in parallel over the nodes. However, in a cloud computing environment, failure is the norm, and nodes may be upgraded, replaced, and added in the system. Files can also be dynamically created, deleted, and appended. This results in load imbalance in a distributed file system; that is, the file chunks are not distributed as uniformly as possible among the nodes. Emerging distributed file systems in production systems strongly depend on a central node for chunk reallocation. This dependence is clearly inadequate in a large-scale, failure-prone environment because the central load balancer is put under considerable workload that is linearly scaled with the system size, and may thus become the performance bottleneck and the single point of failure. In this paper, a fully distributed load rebalancing algorithm is presented to cope with the load imbalance problem. Our algorithm is compared against a centralized approach in a production system and a competing distributed solution presented in the literature. The simulation results indicate that our proposal is comparable with the existing centralized approach and considerably outperforms the prior distributed algorithm in terms of load imbalance factor, movement cost, and algorithmic overhead. The performance of our proposal implemented in the Hadoop distributed file system is further investigated in a cluster environment.

The evolution of massively parallel supercomputers make palpable two issues in particular: the load imbalance and the poor management of data locality in applications. Thus, with the increase of the number of cores and the drastic decrease of amount of memory per core, the large performance needs imply to particularly take care of the load-balancing and as much as possible of the locality of data. One mean to take into account this locality issue relies on the placement of the processing entities and load balancing techniques are relevant in order to improve application performance. With large-scale platforms in mind, we developed a hierarchical and distributed algorithm which aim is to perform a topology-aware load balancing tailored for Charm++ applications. This algorithm is based on both LibTopoMap for the network awareness aspects and on TREEMATCH to determine a relevant placement of the processing entities. We show that the proposed algorithm improves the overall execution time in both the cases of real applications and a synthetic benchmark as well. For this last experiment, we show a scalability up to one millions processing entities.

Virtualization has been shown to be an attractive path to increase overall system resource utilization. The use of live virtual machine (VM) migration has enabled more effective sharing of system resources across multiple physical servers, resulting in an increase in overall performance. Live VM migration can be used to load balance virtualized clusters. To drive live migration, we need to be able to measure the current load imbalance. Further, we also need to accurately predict the resulting load imbalance produced by any migration.In this paper we present a new metric that captures the load of the physical servers and is a function of the resident VMs. This metric will be used to measure load imbalance and construct a load-balancing VM migration framework. The algorithm for balancing the load of virtualized enterprise servers follows a greedy approach, inductively predicting which VM migration will yield the greatest improvement of the imbalance metric in a particular step. We compare our algorithm to the leading commercially available load balancing solution - VMware's Distributed Resource Scheduler (DRS). Our results show that when we are able to accurately measure system imbalance, we can also predict future system state. We find that we can outperform DRS and improve performance up to 5%. Our results show that our approach does not impose additional performance impact and is comparable to the virtual machine monitor overhead.

ABSTRACTMesh adaption is a powerful tool for efficient unstructured-grid computations but causes load imbalance among proces-sors on a parallel machine. We present a novel method todynamically balance the processor workloads with a globalview. This paper presents, for the first time, the imple-mentation and integration of all major components withinour dynamic load balancing strategy for adaptive grid cal-culations. Mesh adaption, repartitioning, processor assign-ment, and remapping are critical components of the frame-work that must be accomplished rapidly and efficiently soas not to cause a significant overhead to the numerical sim-ulation. Previous results indicated that mesh repartitioningand data remapping are potential bottlenecks for perform-ing large-scale scientific calculations. We resolve these issuesand demonstrate that our framework remains viable on alarge number of processors.1 INTRODUCTIONDynamic mesh adaption on unstructured grids is a power-ful tool for computing unsteady three-dimensional problemsthat require grid modifications to efficiently resolve solu-tion features. By locally refining and coarsening the meshto capture flowfield phenomena of interest, such proceduresmake standard computational methods more cost effective.Highly refined meshes are required to accurately captureshock waves, contact discontinuities, vortices, and shear lay-ers. Local mesh adaption provides the opportunity to obtainsolutions that are comparable to those obtained on globally-refined grids but at a much lower cost.Unfortunately, the adaptive solution of unsteady prob-lems causes load imbalance among processors on a paral-lel machine. This is because the computational intensity isboth space and time dependent. An efficient parallel imple-mentation of such methods is extremely difficult to achieve,primarily because of the dynamically-changing nonuniformgrid. Various methods on dynamic load balancing have beenreported to date [5-9,11-14,16-18,24-26]; however, most ofthem either lack a global view of loads across processors orProceedings of the 9th ACM Symposium on Parallel Al-gorithms and Architectures, Newport, Rhode Island, June22-25, 1997.do not apply their techniques to realistic large-scale appli-cations.Figure 1 depicts our framework for parallel adaptive flowcomputation. It consists of a flow solver and a mesh adap-tor, with a partitioner and a remapper that load balancesand redistributes the computational mesh when necessary.Our goal is to build a portable system for efficiently per-forming adaptive large-scale flow calculations in a parallelmessage-passing environment. The mesh is first partitionedand mapped among the available processors. The flow solverthen runs for several iterations, updating solution variables.Once an acceptable solution is obtained, the mesh adaptionprocedure is invoked. It first targets edges for coarseningand refinement based on an error indicator computed fromthe flow solution. The old mesh is then coarsened, resultingin a smaller grid. Since edges have already been marked forrefinement, it is possible to exactly predict the new meshbefore actually performing the refinement step. Programcontrol is thus passed to the load balancer at this time. Aquick evaluation step determines if the new mesh will be sounbalanced as to warrant a repartitioning. If the currentpartitions will remain adequately load balanced, control mpassed back to the subdivision phase of the mesh adaptor.Otherwise, a repartitioning procedure is used to divide thenew mesh into subgrids. The new partitions are then reas-signed to the processors in a way that minimizes the costof data movement. If the remapping cost is less than thecomputational gain that would be achieved with balancedpartitions, all necessary data is appropriately redistributed.Otherwise, the new partitioning is discarded. The computa-tional mesh is then actually refined and the flow calculationis restarted.Notice from the framework in Fig. 1 that splitting themesh refinement step into two distinct phases of edge mark-ing and mesh subdivision allows the subdivision phase tooperate in a more load balanced fashion. In addition, sincedata remapping is performed before the mesh grows in sizedue to refinement, a smaller volume of data is moved. Thiscan lead to a potentially significant savings in the redistribu-tion cost. The load balancer also balances the computationalload for the flow solver while reducing the runtime commu-nication. This is important because flow solvers are usuallyseveral times more expensive than mesh adaptors. In anycase, it is obvious that mesh adaption, repartitioning, pro-cessor assignment, and remapping are critical componentsof the framework and must be accomplished rapidly and ef-ficiently so as not to cause a significant overhead to the flowcomputation.

This paper presents the use of asynchronous algorithms to improve the detection performance of a coded servo channel in the presence of radial phase incoherence. The paper compares the performance of a coded servo system with both the synchronous and asynchronous algorithms. Simulation results are presented under varying amounts of radial incoherence and frequency offset errors.

Routing is probably the most time-consuming stage in the FPGA compilation flow. While parallelization has the potential to reduce the routing time, load imbalance still arises in parallel routing process, resulting in the degradations of speedup and quality of results. Furthermore, when scaling the number of processing cores, load imbalance will become more severe in parallel routing. In this paper, we explore load balance-aware parallel routing to obtain significant speedup. To study load balance problem, we implement a basic MPI-based parallel routing framework on multi-core distributed-memory platform. This framework provides two dynamic partitioning algorithms and both of them can automatically maintain the load balance of parallel routing at runtime. The first is the heuristic algorithm and it is able to assign more nets to fast MPI processes by removing an equal number of nets from slow MPI processes. The second algorithm selects the re-partitioning of the nets for all the MPI processes at each iteration, further leading to a good load balance during parallel routing. The key idea of these two algorithms is to analyze the behaviours of previous iteration to guide the parallelization of next iteration. Results show that the effectiveness and efficiency of the two algorithms in multi-core distributed-memory parallel routing framework. Notably, our parallel router can achieve about 6x speedup on average using 8 MPI processes, comparing to the serial router. This is a 1.2x improvement over a state-of-the-art multi-core parallel router.

A high call blocking rate is a consequence of an inefficient utilization of system resources, which is often caused by a load imbalance in the network. Load imbalances are common in wireless networks with a large number of cellular users. This paper investigates a load-balancing scheme for mobile networks that optimizes cellular performance with constraints of physical resource limits and users quality of service demands. In order to efficiently utilize the system resources, an intelligent distributed antenna system (IDAS) fed by a multi base transceiver station (BTS) has the ability to distribute the system resources over a given geographic area. To enable load balancing among distributed antenna modules we dynamically allocate the remote antenna modules to the BTSs using an intelligent algorithm. A self-optimizing network for an IDAS is formulated as an integer based linear constrained optimization problem, which tries to balance the load among the BTSs. A discrete particle swarm optimization (DPSO) algorithm as an evolutionary algorithm is proposed to solve the optimization problem. The computational results of the DPSO algorithm demonstrate optimum performance for small-scale networks and near-optimum performance for large-scale networks. The DPSO algorithm is faster with marginally less complexity than an exhaustive search algorithm.

This paper targets at evaluating different mechanisms for providing Inter-Frequency (IF) Load Balancing (LB) in advanced multi-layer deployments. In particular, the performance of IF mobility management based on signal quality measurements is compared against a load-dependent Traffic Steering (TS) framework that triggers IF events only if load imbalance is detected. To evaluate the joint interaction of the aforementioned schemes with more advanced LB features, system level simulations have been conducted with and without assuming Carrier Aggregation (CA) capable terminals. Results have shown that although the standardized signal quality-based IF mobility management can maximize data rates, the developed framework reduces significantly the signaling overhead, at the cost of slight User Equipment (UE) throughput degradation. Nevertheless, the observed trade-off almost disappears if CA is enabled, while similar signaling gains can be achieved.

Load-balancing routing in software defined networks with multiple controllers

GPU-accelerated relational query execution engines have parallelized the execution of a pipeline, a sequence of operators. For the parallelization, the engines evenly partition the tuples in a table that will be scanned by the pipeline's first operator (a scan), and each thread executes the pipeline for the tuples in a partition. However, this approach leads to load imbalances since an operator returns a varying number of output tuples per input tuple, particularly under non-uniform data distributions such as skewed join key values. The load imbalances are classified into intra- and inter-warp load imbalances (intra-WLIs and inter-WLIs) since 1) threads are grouped into warps and 2) every thread in a warp evaluates the same operator for an input tuple concurrently following a single-instruction-multiple-thread manner. In contrast, threads in different warps can evaluate different operators concurrently. Although load balancing techniques have been proposed, however, they fail to solve the load imbalances on various workloads. In this paper, we propose a query execution engine, Themis, named after the deity of fairness, which symbolizes balanced workloads within our context. Themis minimizes intra-WLIs and inter-WLIs across various workloads. First, Themis minimizes intra-WLIs by redistributing tuples between the threads in a warp and making the threads evaluate an operator only when all of them hold inputs. Second, Themis mitigates the inter-WLIs by redistributing the tuples of warps with heavy workloads to idle warps. To check whether a warp's workload is heavy, we propose a method to approximate the sizes of warps' workloads. Based on these approximations, Themis adaptively adjusts the threshold for determining a warp's workload as heavy. In a recent benchmark JCC-H, which introduces skewed join key distributions to TPC-H, Themis significantly alleviates the inter-WLIs and intra-WLIs, outperforming the runner-up by up to 379x.

Fine-Grained Parallel and Distributed Spatial Stochastic Simulation of Biological Reactions

This paper describes under frequency based islanding scheme of the Rawatbhatta Atomic Power Plant (RAPP) situated in Rajasthan. The islanding power system consists of seven generators having total 780 MW generation capacity. On the basis of simulation studies, under frequency based islanding scheme have been designed. Under frequency based islanding relays are placed at interconnecting points. Further under frequency based load shedding relays are also placed in the island to control the island frequency in the event of generation and load imbalance. Simulation results are presented up to 10% overload of Island. Simulation studies have been carried out using Mi-Power software. Frequency response of Islanded System, generators active & reactive power output, generators terminal voltage, response of governors & AVRs have been plotted. Simulation studies indicate that with proper setting of under frequency load shedding relays Island is successfully formed and stabilized at acceptable frequency.

The status quo of enterprises' load imbalance in most clusters makes the economical function of the clusters far under the level which them should have. In order to improve the function, it is necessary to carry out working procedures, parts, components and products' order allocation in clusters. At present, there is no an effective and mature method of order allocation for enterprises in a cluster, and enterprise being overload is always because of bottleneck working procedures are overload. Therefore based on working procedures outsourcing, the author puts forward an outsourcing process model for enterprises in a cluster-collaborative order obtaining model (COOM). The model includes two sub-models. First one is outsourcing decision-making, and the second is order allocation process model. UCM is used as describing language of the model. The first sub-model is used for enterprise to decide the outsourcing types and volume of each type. The second is used for enterprise to select proper outsourcers and determine the order allocation volumes. The "information match" is used as the decision-making method for order allocation. In the end, a case is used to prove the feasibility of the model.

In the present cloud computing environment, the scheduling approaches for VM (Virtual Machine) resources only focus on the current state of the entire system. Most often they fail to consider the system variation and historical behavioral data which causes system load imbalance. To present a better approach for solving the problem of VM resource scheduling in a cloud computing environment, this project demonstrates a genetic algorithm based VM resource scheduling strategy that focuses on system load balancing. The genetic algorithm approach computes the impact in advance, that it will have on the system after the new VM resource is deployed in the system, by utilizing historical data and current state of the system. It then picks up the solution, which will have the least effect on the system. By doing this it ensures the better load balancing and reduces the number of dynamic VM migrations. The approach presented in this project solves the problem of load imbalance and high migration costs. Usually load imbalance and high number of VM migrations occur if the scheduling is performed using the traditional algorithms.