Load balancing arises as a fundamental problem, underlying the dimensioning and operation of many computing and communication systems, such as job routing in data center clusters, multipath communication, Big Data and queueing systems. In essence, the decision-making agent maps each arriving job to one of the possibly heterogeneous servers while aiming at an optimization goal such as load balancing, low average delay or low loss rate. One main difficulty in finding optimal load balancing policies here is that the agent only partially observes the impact of its decisions, e.g., through the delayed acknowledgements of the served jobs. In this paper, we provide a partially observable (PO) model that captures the load balancing decisions in parallel buffered systems under limited information of delayed acknowledgements. We present a simulation model for this PO system to find a load balancing policy in real-time using a scalable Monte Carlo tree search algorithm. We numerically show that the resulting policy outperforms other limited information load balancing strategies such as variants of Join-the-Most-Observations and has comparable performance to full information strategies like: Join-the-Shortest-Queue, Join-the-Shortest-Queue(d) and Shortest-Expected-Delay. Finally, we show that our approach can optimise the real-time parallel processing by using network data provided by Kaggle.
Communication load balancing aims to balance the load between different available resources, and thus improve the quality of service for network systems. After formulating the load balancing (LB) as a Markov decision process problem, reinforcement learning (RL) has recently proven effective in addressing the LB problem. To leverage the benefits of classical RL for load balancing, however, we need an explicit reward definition. Engineering this reward function is challenging, because it involves the need for expert knowledge and there lacks a general consensus on the form of an optimal reward f…
Recent years have seen a great increase in the capacity and parallel processing power of data centers and cloud services. To fully utilize the said distributed systems, optimal load balancing for parallel queuing architectures must be realized. Existing state-of-the-art solutions fail to consider the effect of communication delays on the behaviour of very large systems with many clients. In this work, we consider a multi-agent load balancing system, with delayed information, consisting of many clients (load balancers) and many parallel queues. In order to obtain a tractable solution, we model …
Load balancing algorithms play critical roles in systems where the workload has to be distributed across multiple resources, such as cores in multiprocessor system, computers in distributed computing, and network links. In this paper, we study and evaluate four load balancing methods: random, round robin, shortest queue, and shortest queue with stale load information. We build a simulation model and compare mean delay of the systems for the load balancing methods. We also provide a method to improve shortest queue with stale load information load balancing. A performance analysis for the impro…
This paper investigates the network load balancing problem in data centers (DCs) where multiple load balancers (LBs) are deployed, using the multi-agent reinforcement learning (MARL) framework. The challenges of this problem consist of the heterogeneous processing architecture and dynamic environments, as well as limited and partial observability of each LB agent in distributed networking systems, which can largely degrade the performance of in-production load balancing algorithms in real-world setups. Centralised-training-decentralised-execution (CTDE) RL scheme has been proposed to improve M…
Parallel iterative applications often suffer from load imbalance, one of the most critical performance degradation factors. Hence, load balancing techniques are used to distribute the workload evenly to maximize performance. A key challenge is to know \textit{when} to use load balancing techniques. In general, this is done through load balancing criteria, which trigger load balancing based on runtime application data and/or user-defined information. In the first part of this paper, we introduce a novel, automatic load balancing criterion derived from a simple mathematical model. In the second …
In this paper, we derive and investigate approaches to dynamically load balance a distributed task parallel application software. The load balancing strategy is based on task migration. Busy processes export parts of their ready task queue to idle processes. Idle--busy pairs of processes find each other through a random search process that succeeds within a few steps with high probability. We evaluate the load balancing approach for a block Cholesky factorization implementation and observe a reduction in execution time on the order of 5\% in the selected test cases.
We propose a GPU fine-grained load-balancing abstraction that decouples load balancing from work processing and aims to support both static and dynamic schedules with a programmable interface to implement new load-balancing schedules. Prior to our work, the only way to unleash the GPU's potential on irregular problems has been to workload-balance through application-specific, tightly coupled load-balancing techniques. With our open-source framework for load-balancing, we hope to improve programmers' productivity when developing irregular-parallel algorithms on the GPU, and also improve the ove…
Data load balancing is a challenging task in the P2P systems. Distributed hash table (DHT) abstraction, heterogeneous nodes, and non uniform distribution of objects are the reasons to cause load imbalance in structured P2P overlay networks. Previous works solved the load balancing problem by assuming the homogeneous capabilities of nodes, unawareness of the link latency during transferring load, and imposing logical structures to collect and reassign load. We propose a distributed load balancing algorithm with the topology awareness feature by using the concept of virtual servers. In our appro…
This document presents an analysis of different load balance strategies for a Plasma physics code that models high energy particle beams with PIC method. A comparison of different load balancing algorithms is given: static or dynamic ones. Lagrangian and Eulerian partitioning techniques have been investigated.
A parallel computer system is a collection of processing elements that communicate and cooperate to solve large computational problems efficiently. To achieve this, at first the large computational problem is partitioned into several tasks with different work-loads and then are assigned to the different processing elements for computation. Distribution of the work load is known as Load Balancing. An appropriate distribution of work-loads across the various processing elements is very important as disproportional workloads can eliminate the performance benefit of parallelizing the job. Hence, l…
Cloud Computing is a new trend emerging in IT environment with huge requirements of infrastructure and resources. Load Balancing is an important aspect of cloud computing environment. Efficient load balancing scheme ensures efficient resource utilization by provisioning of resources to cloud users on demand basis in pay as you say manner. Load Balancing may even support prioritizing users by applying appropriate scheduling criteria. This paper presents various load balancing schemes in different cloud environment based on requirements specified in Service Level Agreement (SLA).
In cloud environments, load balancing task scheduling is an important issue that directly affects resource utilization. Unquestionably, load balancing scheduling is a serious aspect that must be considered in the cloud research field due to the significant impact on both the back end and front end. Whenever an effective load balance has been achieved in the cloud, then good resource utilization will also be achieved. An effective load balance means distributing the submitted workload over cloud VMs in a balanced way, leading to high resource utilization and high user satisfaction. In this pape…
This paper presents the network load balancing problem, a challenging real-world task for multi-agent reinforcement learning (MARL) methods. Traditional heuristic solutions like Weighted-Cost Multi-Path (WCMP) and Local Shortest Queue (LSQ) are less flexible to the changing workload distributions and arrival rates, with a poor balance among multiple load balancers. The cooperative network load balancing task is formulated as a Dec-POMDP problem, which naturally induces the MARL methods. To bridge the reality gap for applying learning-based methods, all methods are directly trained and evaluate…
In parallel computing, a problem is divided into a set of smaller tasks that are distributed across multiple processing elements. Balancing the load of the processing elements is key to achieving good performance and scalability. If the computational costs of the individual tasks vary over time in an unpredictable way, dynamic load balancing aims at migrating them between processing elements so as to maintain load balance. During dynamic load balancing, the tasks amount to indivisible work packets with a real-valued cost. For this case of indivisible, real- valued loads, we analyze the balanci…
Effective utilization of Multiple-Instruction-Multiple-Data (MIMD) parallel computers requires the application of good load balancing techniques. In this paper we show that heuristics derived from observation of complex natural systems, such as the mammalian immune system, can lead to effective load balancing strategies. In particular, the immune system processes of regulation, suppression, tolerance, and memory are seen to be powerful load balancing mechanisms. We provide a detailed example of our approach applied to parallelization of an image processing task, that of extracting the circuit …
The classification of the most used load balancing algorithms in distributed systems (including cloud technology, cluster systems, grid systems) is described. Comparative analysis of types of the load balancing algorithms is conducted in accordance with the classification, the advantages and drawbags of each type of the algorithms are shown. Performance indicators characterizing each algorithm are indicated.
In the load balancing problem, each node in a network is assigned a load, and the goal is to equally distribute the loads among the nodes, by preforming local load exchanges. While load balancing was extensively studied in static networks, only recently a load balancing algorithm for dynamic networks with a bounded convergence time was presented. In this paper, we further study the time complexity of load balancing in the context of dynamic networks. First, we show that randomness is not necessary, and present a deterministic algorithm which slightly improves the running time of the previous a…
Load balancing is prevalent in practical application (e.g., web) deployments seen today. One such load balancer, HAProxy, remains relevant as an open-source, easy-to-use system. In the context of web systems, the load balancer tier possesses significant influence over system performance and the incurred cost, which is decisive for cloud-based deployments. Therefore, it is imperative to properly tune the load balancer configuration and get the most performance out of the existing resources. In this technical report, we first introduce the HAProxy architecture and its load balancing methods. The…
This work studies distributed algorithms for locally optimal load-balancing: We are given a graph of maximum degree $Δ$, and each node has up to $L$ units of load. The task is to distribute the load more evenly so that the loads of adjacent nodes differ by at most $1$. If the graph is a path ($Δ= 2$), it is easy to solve the fractional version of the problem in $O(L)$ communication rounds, independently of the number of nodes. We show that this is tight, and we show that it is possible to solve also the discrete version of the problem in $O(L)$ rounds in paths. For the general case ($Δ> 2$), w…
Recently, fog computing has been introduced as a modern distributed paradigm and complement to cloud computing to provide services. Fog system extends storing and computing to the edge of the network, which can solve the problem about service computing of the delay-sensitive applications remarkably besides enabling the location awareness and mobility support. Load balancing is an important aspect of fog networks that avoids a situation with some under-loaded or overloaded fog nodes. Quality of Service (QoS) parameters such as resource utilization, throughput, cost, response time, performance, …
In this paper, we introduce DLB, a Deep Learning based load Balancing mechanism, to effectively address the data skew problem. The key idea of DLB is to replace hash functions in the load balancing mechanisms with deep learning models, which are trained to be able to map different distributions of workloads and data to the servers in a uniform manner. We implemented DLB and deployed it on a practical Cloud environment using CloudSim. Experimental results using both synthetic and real-world data sets show that compared with traditional hash function-based load balancing methods, DLB is able to …
The effectiveness and scalability of MapReduce-based implementations of complex data-intensive tasks depend on an even redistribution of data between map and reduce tasks. In the presence of skewed data, sophisticated redistribution approaches thus become necessary to achieve load balancing among all reduce tasks to be executed in parallel. For the complex problem of entity resolution, we propose and evaluate two approaches for such skew handling and load balancing. The approaches support blocking techniques to reduce the search space of entity resolution, utilize a preprocessing MapReduce job…
Load balancing is vital for the efficient and long-term operation of cloud data centers. With virtualization, post (reactive) migration of virtual machines after allocation is the traditional way for load balancing and consolidation. However, reactive migration is not easy to obtain predefined load balance objectives and may interrupt services and bring instability. Therefore, we provide a new approach, called Prepartition, for load balancing. It partitions a VM request into a few sub-requests sequentially with start time, end time and capacity demands, and treats each sub-request as a regular…
With technological advancements and constant changes of Internet, cloud computing has been today's trend. With the lower cost and convenience of cloud computing services, users have increasingly put their Web resources and information in the cloud environment. The availability and reliability of the client systems will become increasingly important. Today cloud applications slightest interruption, the impact will be significant for users. It is an important issue that how to ensure reliability and stability of the cloud sites. Load balancing would be one good solution. This paper presents a fr…
The use of under-utilized Internet resources is widely recognized as a viable form of high performance computing. Sustained processing power of roughly 40T FLOPS using 4 million volunteered Internet hosts has been reported for embarrassingly parallel problems. At the same time, peer-to-peer (P2P) file sharing networks, with more than 50 million participants, have demonstrated the capacity for scale in distributed systems. This paper contributes a study of load balancing techniques for a general class of loosely-synchronous parallel algorithms when executed over a P2P network. We show that dece…
We present a SNN simulator which scales to millions of neurons, billions of synapses, and 8 GPUs. This is made possible by 1) a novel, cache-aware spike transmission algorithm 2) a model parallel multi-GPU distribution scheme and 3) a static, yet very effective load balancing strategy. The simulator further features an easy to use API and the ability to create custom models. We compare the proposed simulator against two state of the art ones on a series of benchmarks using three well-established models. We find that our simulator is faster, consumes less memory, and scales linearly with the nu…
CMT-nek is a new scientific application for performing high fidelity predictive simulations of particle laden explosively dispersed turbulent flows. CMT-nek involves detailed simulations, is compute intensive and is targeted to be deployed on exascale platforms. The moving particles are the main source of load imbalance as the application is executed on parallel processors. In a demonstration problem, all the particles are initially in a closed container until a detonation occurs and the particles move apart. If all processors get an equal share of the fluid domain, then only some of the proce…
Nowadays, multiprocessing is mainstream with exponentially increasing number of processors. Load balancing is, therefore, a critical operation for the efficient execution of parallel algorithms. In this paper we consider the fundamental class of tree-based algorithms that are notoriously irregular, and hard to load-balance with existing static techniques. We propose a hybrid load balancing method using the utility of statistical random sampling in estimating the tree depth and node count distributions to uniformly partition an input tree. To conduct an initial performance study, we implemented…
Nowadays, more and more increasingly hard computations are performed in challenging fields like weather forecasting, oil and gas exploration, and cryptanalysis. Many of such computations can be implemented using a computer cluster with a large number of servers. Incoming computation requests are then, via a so-called load balancing policy, distributed over the servers to ensure optimal performance. Additionally, being able to switch-off some servers during low period of workload, gives potential to reduced energy consumption. Therefore, load balancing forms, albeit indirectly, a trade-off betw…
Fine-grained workload and resource balancing is the key to high performance for regular and irregular computations on the GPUs. In this dissertation, we conduct an extensive survey of existing load-balancing techniques to build an abstraction that addresses the difficulty of scheduling computations on the GPU. We propose a GPU fine-grained load-balancing abstraction that decouples load balancing from work processing and aims to support both static and dynamic schedules with a programmable interface to implement new load-balancing schedules. Prior to our work, the only way to unleash the GPU's …
Tokyo Institute of Technology（东京工业大学）；The University of Tokyo（东京大学）；IBM Watson Research Center（IBM沃森研究中心）；University College Dublin（都柏林大学学院）
We present a scalable parallel solver for numerical constraint satisfaction problems (NCSPs). Our parallelization scheme consists of homogeneous worker solvers, each of which runs on an available core and communicates with others via the global load balancing (GLB) method. The parallel solver is implemented with X10 that provides an implementation of GLB as a library. In experiments, several NCSPs from the literature were solved and attained up to 516-fold speedup using 600 cores of the TSUBAME2.5 supercomputer.
Modern high load applications store data using multiple database instances. Such an architecture requires data consistency, and it is important to ensure even distribution of data among nodes. Load balancing is used to achieve these goals. Hashing is the backbone of virtually all load balancing systems. Since the introduction of classic Consistent Hashing, many algorithms have been devised for this purpose. One of the purposes of the load balancer is to ensure storage cluster scalability. It is crucial for the performance of the whole system to transfer as few data records as possible during n…
Big data analytics in cloud environments introduces challenges such as real-time load balancing besides security, privacy, and energy efficiency. In this paper, we propose a novel load balancing algorithm in cloud environments that performs resource allocation and task scheduling efficiently. The proposed load balancer reduces the execution response time in big data applications performed on clouds. Scheduling, in general, is an NP-hard problem. In our proposed algorithm, we provide solutions to reduce the search area that leads to reduced complexity of the load balancing. We recommend two mat…
Maintaining computational load balance is important to the performant behavior of codes which operate under a distributed computing model. This is especially true for GPU architectures, which can suffer from memory oversubscription if improperly load balanced. We present enhancements to traditional load balancing approaches and explicitly target GPU architectures, exploring the resulting performance. A key component of our enhancements is the introduction of several GPU-amenable strategies for assessing compute work. These strategies are implemented and benchmarked to find the most optimal dat…
For the parallel computation of partial differential equations, one key is the grid partitioning. It requires that each process owns the same amount of computations, and also, the partitioning quality should be proper to reduce the communications among processes. When calculating the partial differential equations using adaptive finite element methods, the grid and the basis functions adjust in each iteration, which introduce load balancing issues. The grid should be redistributed dynamically. This paper studies dynamic load balancing algorithms and the implementation on the adaptive finite el…
This paper studies the problem of load-balancing the demand for content in a peer-to-peer network across heterogeneous peer nodes that hold replicas of the content. Previous decentralized load balancing techniques in distributed systems base their decisions on periodic updates containing information about load or available capacity observed at the serving entities. We show that these techniques do not work well in the peer-to-peer context; either they do not address peer node heterogeneity, or they suffer from significant load oscillations. We propose a new decentralized algorithm, Max-Cap, ba…
In this paper we study dynamic averaging load balancing on general graphs. We consider infinite time and dynamic processes, where in every step new load items are assigned to randomly chosen nodes. A matching is chosen, and the load is averaged over the edges of that matching. We analyze the discrete case where load items are indivisible, moreover our results also carry over to the continuous case where load items can be split arbitrarily. For the choice of the matchings we consider three different models, random matchings of linear size, random matchings containing only single edges, and dete…
Heavy traffic analysis for load balancing policies has relied heavily on the condition of state-space collapse onto a single-dimensional line in previous works. In this paper, via Lyapunov-drift analysis, we rigorously prove that even under a multi-dimensional state-space collapse, steady-state heavy-traffic delay optimality can still be achieved for a general load balancing system. This result directly implies that achieving steady-state heavy-traffic delay optimality simply requires that no server is idling while others are busy at heavy loads, thus complementing and extending the result obt…
This article proposes a design and implementation of a low cost two-tier architecture model for high availability cluster combined with load-balancing and shared storage technology to achieve desired scale of three-tier architecture for application load balancing e.g. web servers. The research work proposes a design that physically omits Network File System (NFS) server nodes and implements NFS server functionalities within the cluster nodes, through Red Hat Cluster Suite (RHCS) with High Availability (HA) proxy load balancing technologies. In order to achieve a low-cost implementation in term…
The problem of minimizing mean response time of generic jobs submitted to a heterogenous distributed computer systems is considered in this paper. A static load balancing strategy, in which decision of redistribution of loads does not depend on the state of the system, is used for this purpose. The article is closely related to a previous article on the same topic. The present article points out number of inconsistencies in the previous article, provides a new formulation, and discusses the impact of new findings, based on the improved formulation, on the results of the previous article.
We study the problem of load balancing in distributed stream processing engines, which is exacerbated in the presence of skew. We introduce Partial Key Grouping (PKG), a new stream partitioning scheme that adapts the classical "power of two choices" to a distributed streaming setting by leveraging two novel techniques: key splitting and local load estimation. In so doing, it achieves better load balancing than key grouping while being more scalable than shuffle grouping. We test PKG on several large datasets, both real-world and synthetic. Compared to standard hashing, PKG reduces the load imb…
Load balancing is the process of improving the Performance of a parallel and distributed system through is distribution of load among the processors [1-2]. Most of the previous work in load balancing and distributed decision making in general, do not effectively take into account the uncertainty and inconsistency in state information but in fuzzy logic, we have advantage of using crisps inputs. In this paper, we present a new approach for implementing dynamic load balancing algorithm with fuzzy logic, which can face to uncertainty and inconsistency of previous algorithms, further more our algo…
With the rapid development in wide area networks and low cost, powerful computational resources, grid computing has gained its popularity. With the advent of grid computing, space limitations of conventional distributed systems can be overcome and underutilized computing resources at different locations around the world can be put to distributed jobs. Workload and resource management is the main key grid services at the service level of grid infrastructures, out of which load balancing in the main concern for grid developers. It has been found that load is the major problem which server faces,…
We present an open-source topology-aware hierarchical unstructured mesh partitioning and load-balancing tool TreePart. The framework provides powerful abstractions to automatically detect and build hierarchical MPI topology resembling the hardware at runtime. Using this information it intelligently chooses between shared and distributed parallel algorithms for partitioning and load-balancing. It provides a range of partitioning methods by interfacing with existing shared and distributed memory parallel partitioning libraries. It provides powerful and scalable abstractions like one-sided distri…
We consider the load balancing problem in large-scale heterogeneous systems with multiple dispatchers. We introduce a general framework called Local-Estimation-Driven (LED). Under this framework, each dispatcher keeps local (possibly outdated) estimates of queue lengths for all the servers, and the dispatching decision is made purely based on these local estimates. The local estimates are updated via infrequent communications between dispatchers and servers. We derive sufficient conditions for LED policies to achieve throughput optimality and delay optimality in heavy-traffic, respectively. Th…
We consider the following load balancing process for $m$ tokens distributed arbitrarily among $n$ nodes connected by a complete graph: In each time step a pair of nodes is selected uniformly at random. Let $\ell_1$ and $\ell_2$ be their respective number of tokens. The two nodes exchange tokens such that they have $\lceil(\ell_1 + \ell_2)/2\rceil$ and $\lfloor(\ell_1 + \ell_2)/2\rfloor$ tokens, respectively. We provide a simple analysis showing that this process reaches almost perfect balance within $O(n\log{n} + n \logΔ)$ steps, where $Δ$ is the maximal initial load difference between any two…
E-science applications may require huge amounts of data and high processing power where grid infrastructures are very suitable for meeting these requirements. The load distribution in a grid may vary leading to the bottlenecks and overloaded sites. We describe a hierarchical dynamic load balancing protocol for Grids. The Grid consists of clusters and each cluster is represented by a coordinator. Each coordinator first attempts to balance the load in its cluster and if this fails, communicates with the other coordinators to perform transfer or reception of load. This process is repeated periodi…
Current-day data centers and high-volume cloud services employ a broad set of heterogeneous servers. In such settings, client requests typically arrive at multiple entry points, and dispatching them to servers is an urgent distributed systems problem. This paper presents an efficient solution to the load balancing problem in such systems that improves on and overcomes problems of previous solutions. The load balancing problem is formulated as a stochastic optimization problem, and an efficient algorithmic solution is obtained based on a subtle mathematical analysis of the problem. Finally, ext…
Particle advection is the approach for extraction of integral curves from vector fields. Efficient parallelization of particle advection is a challenging task due to the problem of load imbalance, in which processes are assigned unequal workloads, causing some of them to idle as the others are performing compute. Various approaches to load balancing exist, yet they all involve trade-offs such as increased inter-process communication, or the need for central control structures. In this work, we present two local load balancing methods for particle advection based on the family of diffusive load…
Equation systems resulting from a p-version FEM discretisation typically require a special treatment as iterative solvers are not very efficient here. Applying hierarchical concepts based on a nested dissection approach allow for both the design of sophisticated solvers as well as for advanced parallelisation strategies. To fully exploit the underlying computing power of parallel systems, dynamic load balancing strategies become an essential component.
