ترغب بنشر مسار تعليمي؟ اضغط هنا

On the Complexity of Load Balancing in Dynamic Networks

84   0   0.0 ( 0 )
 نشر من قبل Ami Paz
 تاريخ النشر 2021
  مجال البحث الهندسة المعلوماتية
والبحث باللغة English




اسأل ChatGPT حول البحث

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 algorithm, at the price of not being matching-based. Then, we consider integral loads, i.e., loads that cannot be split indefinitely, and prove that no matching-based algorithm can have a bounded convergence time for this case. To circumvent both this impossibility result, and a known one for the non-integral case, we apply the method of smoothed analysis, where random perturbations are made over the worst-case choices of network topologies. We show both impossibility results do not hold under this kind of analysis, suggesting that load-balancing in real world systems might be faster than the lower bounds suggest.



قيم البحث

اقرأ أيضاً

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 s ophisticated 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.
Random hashing is a standard method to balance loads among nodes in Peer-to-Peer networks. However, hashing destroys locality properties of object keys, the critical properties to many applications, more specifically, those that require range searchi ng. To preserve a key order while keeping loads balanced, Ganesan, Bawa and Garcia-Molina proposed a load-balancing algorithm that supports both object insertion and deletion that guarantees a ratio of 4.237 between the maximum and minimum loads among nodes in the network using constant amortized costs. However, their algorithm is not straightforward to implement in real networks because it is recursive. Their algorithm mostly uses local operations with global max-min load information. In this work, we present a simple non-recursive algorithm using essentially the same primitive operations as in Ganesan {em et al.}s work. We prove that for insertion and deletion, our algorithm guarantees a constant max-min load ratio of 7.464 with constant amortized costs.
We present a novel framework, called balanced overlay networks (BON), that provides scalable, decentralized load balancing for distributed computing using large-scale pools of heterogeneous computers. Fundamentally, BON encodes the information about each nodes available computational resources in the structure of the links connecting the nodes in the network. This distributed encoding is self-organized, with each node managing its in-degree and local connectivity via random-walk sampling. Assignment of incoming jobs to nodes with the most free resources is also accomplished by sampling the nodes via short random walks. Extensive simulations show that the resulting highly dynamic and self-organized graph structure can efficiently balance computational load throughout large-scale networks. These simulations cover a wide spectrum of cases, including significant heterogeneity in available computing resources and high burstiness in incoming load. We provide analytical results that prove BONs scalability for truly large-scale networks: in particular we show that under certain ideal conditions, the network structure converges to Erdos-Renyi (ER) random graphs; our simulation results, however, show that the algorithm does much better, and the structures seem to approach the ideal case of d-regular random graphs. We also make a connection between highly-loaded BONs and the well-known ball-bin randomized load balancing framework.
Distributed processing across a networked environment suffers from unpredictable behavior of speedup due to heterogeneous nature of the hardware and software in the remote machines. It is challenging to get a better performance from a distributed sys tem by distributing task in an intelligent manner such that the heterogeneous nature of the system do not have any effect on the speedup ratio. This paper introduces homogenization, a technique that distributes and balances the workload in such a manner that the user gets the highest speedup possible from a distributed environment. Along with providing better performance, homogenization is totally transparent to the user and requires no interaction with the system.
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 data collection methodology for in-situ assessment of GPU compute work. For the fully kinetic particle-in-cell code WarpX, which supports MPI+CUDA parallelism, we investigate the performance of the improved dynamic load balancing via a strong scaling-based performance model and show that, for a laser-ion acceleration test problem run with up to 6144 GPUs on Summit, the enhanced dynamic load balancing achieves from 62%--74% (88% when running on 6 GPUs) of the theoretically predicted maximum speedup; for the 96-GPU case, we find that dynamic load balancing improves performance relative to baselines without load balancing (3.8x speedup) and with static load balancing (1.2x speedup). Our results provide important insights into dynamic load balancing and performance assessment, and are particularly relevant in the context of distributed memory applications ran on GPUs.
التعليقات
جاري جلب التعليقات جاري جلب التعليقات
سجل دخول لتتمكن من متابعة معايير البحث التي قمت باختيارها
mircosoft-partner

هل ترغب بارسال اشعارات عن اخر التحديثات في شمرا-اكاديميا