Dynamic Load Balancing in Parallel Execution of Cellular Automata

Giordano, Andrea; Rango, Alessio De; Rongo, Rocco; D'Ambrosio, Donato; Spataro, William

doi:10.1109/tpds.2020.3025102

Cited by 22 publications

(17 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Ref. [9] proposed a dynamic load-balancing solution evaluated using the computational paradigm related to the spatially related Cellular Automata. The major contribution of this research is the development of simple closed-form equations that enable the calculation of the best workload assignment dynamically, intending to assure a perfectly balanced workload distribution throughout the parallel execution.…”

Section: Related Workmentioning

confidence: 99%

A Fog-Cluster Based Load-Balancing Technique

et al. 2022

View full text Add to dashboard Cite

The Internet of Things has recently been a popular topic of study for developing smart homes and smart cities. Most IoT applications are very sensitive to delays, and IoT sensors provide a constant stream of data. The cloud-based IoT services that were first employed suffer from increased latency and inefficient resource use. Fog computing is used to address these issues by moving cloud services closer to the edge in a small-scale, dispersed fashion. Fog computing is quickly gaining popularity as an effective paradigm for providing customers with real-time processing, platforms, and software services. Real-time applications may be supported at a reduced operating cost using an integrated fog-cloud environment that minimizes resources and reduces delays. Load balancing is a critical problem in fog computing because it ensures that the dynamic load is distributed evenly across all fog nodes, avoiding the situation where some nodes are overloaded while others are underloaded. Numerous algorithms have been proposed to accomplish this goal. In this paper, a framework was proposed that contains three subsystems named user subsystem, cloud subsystem, and fog subsystem. The goal of the proposed framework is to decrease bandwidth costs while providing load balancing at the same time. To optimize the use of all the resources in the fog sub-system, a Fog-Cluster-Based Load-Balancing approach along with a refresh period was proposed. The simulation results show that “Fog-Cluster-Based Load Balancing” decreases energy consumption, the number of Virtual Machines (VMs) migrations, and the number of shutdown hosts compared with existing algorithms for the proposed framework.

show abstract

Section: Related Workmentioning

confidence: 99%

A Fog-Cluster Based Load-Balancing Technique

et al. 2022

View full text Add to dashboard Cite

show abstract

“…(1) The evaluation index of node load is single. Different DSs deal with different scenarios [9][10]. Therefore, the factors affecting the load of distributed nodes in the cluster are also different, and it is difficult to directly evaluate the load of nodes through a certain index.…”

Section: Defects Of Traditional Algorithmsmentioning

confidence: 99%

Distributed System Load Balancing Strategy for Converged Workstations

2022

DPS

View full text Add to dashboard Cite

In the era of cloud computing and big data, distributed block storage system has become more and more important with its unique advantages. Load balancing (LB) is an important feature of distributed block storage systems, and it is also one of the hotspots of current distributed block storage research. Based on the existing LB algorithms, researching new and more effective LB techniques is very effective in improving the stability of the system. The main purpose of this paper is to study the LB strategy in the distributed system (DS) of converged workstations. Based on the traditional LB strategy, this paper selects multiple evaluation indicators and establishes a corresponding evaluation function to improve the accuracy of judging the actual load of the node server. According to the situation of different hardware parts, the hierarchical standard of the load state is formulated, and the instantaneous stress test of the improved LB algorithm and the weighted minimum connection algorithm is carried out. The complexity of a LB algorithm. Experiments show that with the increase of the number of request connections, the response time of the system under the two algorithms is longer, but the response time of the improved LB scheduling algorithm has a smaller increase. That is to say, the improved LB algorithm is better than the weighted least connection algorithm in handling the instantaneous pressure of the system.

show abstract

“…In practical distributed systems, many applications are complex, and it is not easy to predetermine the computational and communication costs of tasks, and even the tasks of some applications will vary as the process progresses [16][17]. Therefore, the classification methods described above may not be able to efficiently complete the load balancing task, so we propose a dynamic load balancing strategy.…”

Section: Dynamic Load Balancingmentioning

confidence: 99%

Implementation Model of Dynamic Load Balancing in Distributed System Based on Consistent Hash Algorithm

2021

DPS

View full text Add to dashboard Cite

In the new stage of network computing, the research and application of distributed systems are becoming more and more extensive. Dynamic load balancing is the core technology of network computing. How to improve the performance of dynamic load balancing has always been a research hotspot of network researchers. The purpose of this paper is to implement a dynamic load balancing model based on a consistent hashing algorithm in a distributed system, and propose a consistent hashing algorithm based on virtual nodes. Firstly, the distributed system and distributed object technology are discussed, and then the load balancing algorithms and models are studied in each chapter. By introducing the concepts of system resource utilization and virtual nodes, the performance and adaptability of the consistent load-constrained hashing algorithm are optimized. To see the prediction accuracy more clearly, randomly select the predicted state of one of the bins. The success rate of the one-step prediction model shows that the success rate of the algorithm's prediction model is generally between 0.8 and 0.9, and the hit rate is relatively stable. The success rate is slightly lower than the single-step success rate, generally between 0.3 and 0.6.

show abstract

Dynamic Load Balancing in Parallel Execution of Cellular Automata

Cited by 22 publications

References 31 publications

A Fog-Cluster Based Load-Balancing Technique

A Fog-Cluster Based Load-Balancing Technique

Distributed System Load Balancing Strategy for Converged Workstations

Implementation Model of Dynamic Load Balancing in Distributed System Based on Consistent Hash Algorithm

Contact Info

Product

Resources

About