scholarly journals Enhancing Robustness of Per-Packet Load-Balancing for Fat-Tree

2021 ◽  
Vol 11 (6) ◽  
pp. 2664
Author(s):  
Chansook Lim
Keyword(s):  
Load Balancing ◽  
Queue Length ◽  
Completion Time ◽  
Low Flow ◽  
Network Failure ◽  
Packet Reordering ◽  
Network Utilization ◽  
Queueing Delay ◽  
Simulation Results ◽  
Cost Paths

Fat-tree networks have many equal-cost redundant paths between two hosts. To achieve low flow completion time and high network utilization in fat-tree, there have been many efforts to exploit topological symmetry. For example, packet scatter schemes, which spray packets across all equal-cost paths relying on topological symmetry, work well when there is no failure in networks. However, when symmetry of a network is disturbed due to a network failure, packet scatter schemes may suffer massive packet reordering. In this paper, we propose a new load balancing scheme named LBSP (Load Balancing based on Symmetric Path groups) for fat-trees. LBSP partitions equal-cost paths into equal sized path groups and assigns a path group to each flow so that packets of a flow are forwarded across paths within the selected path group. When a link failure occurs, the flows affected by the failure are assigned an alternative path group which does not contain the failed link. Consequently, packets in one flow can still experience almost the same queueing delay. Simulation results show that LBSP is more robust to network failures compared to the original packet scatter scheme. We also suggest a solution to the queue length differentials between path groups.

Analysis & Minimization of the Effect of Delay on Load Balancing for Efficient Web Server Queueing Model

2014 ◽  
Vol 3 (4) ◽  
pp. 1-16 ◽  
Author(s):  
Harikesh Singh ◽  
Shishir Kumar
Keyword(s):  
Load Balancing ◽  
Waiting Time ◽  
Queue Length ◽  
Completion Time ◽  
Web Server ◽  
Task Completion ◽  
Web Servers ◽  
Task Completion Time ◽  
Average Waiting Time ◽  
The Web

Load balancing applications introduce delays due to load relocation among various web servers and depend upon the design of balancing algorithms and resources required to share in the large and wide applications. The performance of web servers depends upon the efficient sharing of the resources and it can be evaluated by the overall task completion time of the tasks based on the load balancing algorithm. Each load balancing algorithm introduces delay in the task allocation among the web servers, but still improved the performance of web servers dynamically. As a result, the queue-length of web server and average waiting time of tasks decreases with load balancing instants based on zero, deterministic, and random types of delay. In this paper, the effects of delay due to load balancing have been analyzed based on the factors: average queue-length and average waiting time of tasks. In the proposed Ratio Factor Based Delay Model (RFBDM), the above factors are minimized and improved the functioning of the web server system based on the average task completion time of each web server node. Based on the ratio of average task completion time, the average queue-length and average waiting time of the tasks allocated to the web server have been analyzed and simulated with Monte-Carlo simulation. The results of simulation have shown that the effects of delays in terms of average queue-length and average waiting time using proposed model have minimized in comparison to existing delay models of the web servers.

Cloud Computing in Three Phase Scheduling Algorithm Based on Service Quality

2014 ◽  
Vol 687-691 ◽  
pp. 2862-2866
Author(s):  
Yong Long Zhuang ◽  
Xiao Lan Weng ◽  
Xiang He Wei
Keyword(s):  
Cloud Computing ◽  
Load Balancing ◽  
Service Quality ◽  
Network Topology ◽  
Completion Time ◽  
Scheduling Algorithm ◽  
Hierarchical Network ◽  
Three Phase ◽  
Simulation Results ◽  
Minimum Completion Time

Assurance issues for cloud computing’s service quality; propose a hierarchical network topology on the basis of the three-stage scheduling algorithm. The algorithm ensures that each work to be performed can be assigned to the appropriate resources, and can effectively improve the load of each node and reduce the waste of resources. FCFS, SJF, OLB and Min-Min (MM) scheduling algorithm’s simulation results show that the algorithm can not only make the work get the minimum completion time, but also the service node operation achieve the load balancing.

scholarly journals HTPC: heterogeneous traffic-aware partition coding for random packet spraying in data center networks

2021 ◽  
Vol 10 (1) ◽  
Author(s):  
Jiawei Huang ◽  
Shiqi Wang ◽  
Shuping Li ◽  
Shaojun Zou ◽  
Jinbin Hu ◽  
...  
Keyword(s):  
Data Center ◽  
Large Scale ◽  
Network Performance ◽  
Rooted Tree ◽  
Heterogeneous Traffic ◽  
Data Center Networks ◽  
Packet Reordering ◽  
Traffic Characteristics ◽  
Network Utilization ◽  
The Impact

AbstractModern data center networks typically adopt multi-rooted tree topologies such leaf-spine and fat-tree to provide high bisection bandwidth. Load balancing is critical to achieve low latency and high throughput. Although the per-packet schemes such as Random Packet Spraying (RPS) can achieve high network utilization and near-optimal tail latency in symmetric topologies, they are prone to cause significant packet reordering and degrade the network performance. Moreover, some coding-based schemes are proposed to alleviate the problem of packet reordering and loss. Unfortunately, these schemes ignore the traffic characteristics of data center network and cannot achieve good network performance. In this paper, we propose a Heterogeneous Traffic-aware Partition Coding named HTPC to eliminate the impact of packet reordering and improve the performance of short and long flows. HTPC smoothly adjusts the number of redundant packets based on the multi-path congestion information and the traffic characteristics so that the tailing probability of short flows and the timeout probability of long flows can be reduced. Through a series of large-scale NS2 simulations, we demonstrate that HTPC reduces average flow completion time by up to 60% compared with the state-of-the-art mechanisms.

scholarly journals Latency-Optimal Computational Offloading Strategy for Sensitive Tasks in Smart Homes

Sensors ◽  
2021 ◽  
Vol 21 (7) ◽  
pp. 2347
Author(s):  
Yanyan Wang ◽  
Lin Wang ◽  
Ruijuan Zheng ◽  
Xuhui Zhao ◽  
Muhua Liu
Keyword(s):  
Queue Length ◽  
Optimization Problem ◽  
Time Slot ◽  
Smart Homes ◽  
Back Pressure ◽  
Smart Devices ◽  
Smart Device ◽  
Simulation Results ◽  
Computational Offloading ◽  
The Stability

In smart homes, the computational offloading technology of edge cloud computing (ECC) can effectively deal with the large amount of computation generated by smart devices. In this paper, we propose a computational offloading strategy for minimizing delay based on the back-pressure algorithm (BMDCO) to get the offloading decision and the number of tasks that can be offloaded. Specifically, we first construct a system with multiple local smart device task queues and multiple edge processor task queues. Then, we formulate an offloading strategy to minimize the queue length of tasks in each time slot by minimizing the Lyapunov drift optimization problem, so as to realize the stability of queues and improve the offloading performance. In addition, we give a theoretical analysis on the stability of the BMDCO algorithm by deducing the upper bound of all queues in this system. The simulation results show the stability of the proposed algorithm, and demonstrate that the BMDCO algorithm is superior to other alternatives. Compared with other algorithms, this algorithm can effectively reduce the computation delay.

scholarly journals Cost minimization of algorithms for scheduling parallel, single-threaded, heterogeneous, speed-scalable processors

2021 ◽  
Author(s):  
Rashid Khogali
Keyword(s):  
Load Balancing ◽  
Response Time ◽  
Scaling Function ◽  
Unit Price ◽  
Energy Sources ◽  
Speed Scaling ◽  
Computing Architecture ◽  
Performance Metric ◽  
Simulation Results ◽  
Better Than

We synthesize online scheduling algorithms to optimally assign a set of arriving heterogeneous tasks to heterogeneous speed-scalable processors under the single threaded computing architecture. By using dynamic speed-scaling, where each processor's speed is able to dynamically change within hardware and software processing constraints, the goal of our algorithms is to minimize the total financial cost (in dollars) of response time and energy consumption (TCRTEC) of the tasks. In our work, the processors are heterogeneous in that they may differ in their hardware specifications with respect to maximum processing rate, power function parameters and energy sources. Tasks are heterogeneous in terms of computation volume, memory and minimum processing requirements. We also consider that the unit price of response time for each task is heterogeneous because the user may be willing to pay higher/lower unit prices for certain tasks, thereby increasing/decreasing their optimum processing rates. We model the overhead loading time incurred when a task is loaded by a given processor prior to its execution and assume it to be heterogeneous as well. Under the single threaded, single buffered computing architecture, we synthesize the SBDPP algorithm and its two other versions. Its first two versions allow the user to specify the unit price of energy and response time for executing each arriving task. The algorithm's second version extends the functionality of the first by allowing the user or the OS of the computing device to further modify a task's unit price of time or energy in order to achieve a linearly controlled operation point that lies somewhere in the economy-performance mode continuum of a task's execution. The algorithm's third version operates exclusively on the latter. We briefly extend the algorithm and its versions to consider migration, where an unfinished task is paused and resumed on another processor. The SBDPP algorithm is qualitatively compared against its two other versions. The SBDPP dispatcher is analytically shown to perform better than the well known Round Robin dispatcher in terms of the TCRTEC performance metric. Through simulations we deduce a relationship between the arrival rate of tasks, number of processors and response time of tasks. Under the Single threaded, multi-buffered computing architecture we have four contributions that constitute the SMBSPP algorithm. First, we propose a novel task dispatching strategy for assigning the tasks to the processors. Second, we propose a novel preemptive service discipline called Smallest remaining Computation Volume Per unit Price of response Time (SCVPPT) to schedule the tasks on the assigned processor. Third, we propose a dynamic speed-scaling function that explicitly determines the optimum processing rate of each task. Most of the simulations consider both stochastic and deterministic traffic conditions. Our simulation results show that SCVPPT outperforms the two known service disciplines, Shortest Remaining Processing Time (SRPT) and the First Come First Serve (FCFS), in terms of minimizing the TCRTEC performance metric. The results also show that the algorithm's dispatcher drastically outperforms the well known Round Robin dispatcher with cost savings exceeding 100% even when the processors are mildly heterogeneous. Finally, analytical and simulation results show that our speed scaling function performs better than a comparable speed scaling function in current literature. Under a fixed budget of energy, we synthesize the SMBAD algorithm which uses the micro-economic laws of Supply and Demand (LSD) to heuristically adjust the unit price of energy in order to extend battery life and execute more than 50% of tasks on a single processor (under the single threaded, multi buffered computing architecture). By extending all our multiprocessor algorithms to factor independent (battery) energy sources that is associated with each processor, we analytically show that load balancing effects are induced on hetergeneous parallel processors. This happens when the unit price of energy is adjusted by the battery level of each processor in accordance with LSD. Furthermore, we show that a variation of this load balancing effect also occurs when the heterogeneous processors use a single battery as long as they operate at unconstrained processing rates.

scholarly journals Cost minimization of algorithms for scheduling parallel, single-threaded, heterogeneous, speed-scalable processors

2021 ◽  
Author(s):  
Rashid Khogali
Keyword(s):  
Load Balancing ◽  
Response Time ◽  
Scaling Function ◽  
Unit Price ◽  
Energy Sources ◽  
Speed Scaling ◽  
Computing Architecture ◽  
Performance Metric ◽  
Simulation Results ◽  
Better Than

We synthesize online scheduling algorithms to optimally assign a set of arriving heterogeneous tasks to heterogeneous speed-scalable processors under the single threaded computing architecture. By using dynamic speed-scaling, where each processor's speed is able to dynamically change within hardware and software processing constraints, the goal of our algorithms is to minimize the total financial cost (in dollars) of response time and energy consumption (TCRTEC) of the tasks. In our work, the processors are heterogeneous in that they may differ in their hardware specifications with respect to maximum processing rate, power function parameters and energy sources. Tasks are heterogeneous in terms of computation volume, memory and minimum processing requirements. We also consider that the unit price of response time for each task is heterogeneous because the user may be willing to pay higher/lower unit prices for certain tasks, thereby increasing/decreasing their optimum processing rates. We model the overhead loading time incurred when a task is loaded by a given processor prior to its execution and assume it to be heterogeneous as well. Under the single threaded, single buffered computing architecture, we synthesize the SBDPP algorithm and its two other versions. Its first two versions allow the user to specify the unit price of energy and response time for executing each arriving task. The algorithm's second version extends the functionality of the first by allowing the user or the OS of the computing device to further modify a task's unit price of time or energy in order to achieve a linearly controlled operation point that lies somewhere in the economy-performance mode continuum of a task's execution. The algorithm's third version operates exclusively on the latter. We briefly extend the algorithm and its versions to consider migration, where an unfinished task is paused and resumed on another processor. The SBDPP algorithm is qualitatively compared against its two other versions. The SBDPP dispatcher is analytically shown to perform better than the well known Round Robin dispatcher in terms of the TCRTEC performance metric. Through simulations we deduce a relationship between the arrival rate of tasks, number of processors and response time of tasks. Under the Single threaded, multi-buffered computing architecture we have four contributions that constitute the SMBSPP algorithm. First, we propose a novel task dispatching strategy for assigning the tasks to the processors. Second, we propose a novel preemptive service discipline called Smallest remaining Computation Volume Per unit Price of response Time (SCVPPT) to schedule the tasks on the assigned processor. Third, we propose a dynamic speed-scaling function that explicitly determines the optimum processing rate of each task. Most of the simulations consider both stochastic and deterministic traffic conditions. Our simulation results show that SCVPPT outperforms the two known service disciplines, Shortest Remaining Processing Time (SRPT) and the First Come First Serve (FCFS), in terms of minimizing the TCRTEC performance metric. The results also show that the algorithm's dispatcher drastically outperforms the well known Round Robin dispatcher with cost savings exceeding 100% even when the processors are mildly heterogeneous. Finally, analytical and simulation results show that our speed scaling function performs better than a comparable speed scaling function in current literature. Under a fixed budget of energy, we synthesize the SMBAD algorithm which uses the micro-economic laws of Supply and Demand (LSD) to heuristically adjust the unit price of energy in order to extend battery life and execute more than 50% of tasks on a single processor (under the single threaded, multi buffered computing architecture). By extending all our multiprocessor algorithms to factor independent (battery) energy sources that is associated with each processor, we analytically show that load balancing effects are induced on hetergeneous parallel processors. This happens when the unit price of energy is adjusted by the battery level of each processor in accordance with LSD. Furthermore, we show that a variation of this load balancing effect also occurs when the heterogeneous processors use a single battery as long as they operate at unconstrained processing rates.

scholarly journals EVALUATION OF THE ROUTING METIRC W-METRIC USED WITH RPL PRTOCOL IN LLNS

2018 ◽  
Vol 19 (2) ◽  
pp. 80-89
Author(s):  
Rosminazuin Ab. Rahim ◽  
Abdallah Awad ◽  
Aisha Hassan Abdalla Hashim ◽  
ALIZA AINI MD RALIB
Keyword(s):  
Load Balancing ◽  
Low Power ◽  
Routing Protocol ◽  
Queue Length ◽  
Packet Delivery Ratio ◽  
Delivery Ratio ◽  
Performance Parameters ◽  
Lossy Networks ◽  
Promising Alternative ◽  
Packet Delivery

ABSTRACT: The current de-facto routing protocol over Low Power and Lossy Networks (LLN) developed by the IETF Working Group (6LOWPAN), is named as Routing Protocol for Low Power and Lossy networks (RPL). RPL in the network layer faces throughput  challenges due to the potential  large networks, number of nodes, and that  multiple  coexisting applications  will  be  running  in  the  same physical layer.  In this study, a node metric for RPL protocol based on the node’s Queue Backlogs is introduced, which leads to a better throughput performance while maintaining the delay and the ability to use with different network applications. This metric depends on the length of Packet Queue of the nodes with the consideration of other link and node metrics, like ETX or energy usage, leading to better load balancing in the network. To implement and evaluate the proposed metric compared to other RPL metrics, ContikiOS and COOJA simulator are used. Extensive simulations have been carried out in a systematic way resulting in a detailed analysis of the introduced metric namely W-metric, expected transmission count (ETX) and objective function zero (OF0) that uses hop-count as a routing metric. The analysis and comparison are based on five performance parameters, which are throughput, packet delivery ratio (PDR), latency, average queue length, and power consumption. Simulation results show that the introduced W-metric has a good performance compared to other RPL metrics with regards to performance parameters mentioned above. At the same time, the results show that its latency performance is comparable with other RPL routing metrics. In a sample simulation of 500 seconds with 25 nodes and with nodes sending packets periodically to the network root at a rate of 1 packet per 4 seconds, W-metric showed a very efficient throughput of 5.16 kbps, an increase of 8.2% compared to ETX. Results showed that it has a packet delivery ratio of 93.3%, which is higher compared to 83.3% for ETX and 74.2% for OF0. Average queue length of 0.48 packet shows improvement of 15.8% better than ETX. In addition, it exhibits an energy consumption of 5.16 mW which is 2.1% less than ETX. Overall, W-metric appears to be a promising alternative to ETX and OF0 as it selects routes that are more efficient by working on load balancing of the network and by considering the link characteristics. ABSTRAK: Protokol penghalaan de-facto semasa ke atas Rangkaian Kekuatan Rendah dan Lossy yang dibangunkan oleh Kumpulan Kerja IETF (6LOWPAN), dinamakan Protokol Penghalaan untuk Kekuatan Rendah dan Rugi (RPL). RPL dalam lapisan rangkaian menghadapi cabaran throughput berikutan jangkaan rangkaian besar, bilangan nod dan aplikasi berganda bersama akan diproses dalam lapisan fizikal yang sama. Dalam kajian ini, satu metrik nod untuk protokol RPL berdasarkan pada Backend Queue node diperkenalkan, yang membawa kepada prestasi yang lebih baik sambil mengekalkan kelewatan dan keupayaan untuk digunakan dengan aplikasi rangkaian yang berbeza. Metrik ini bergantung pada panjang Packet Queue dari node dengan pertimbangan metrik lain dan nodus lain, seperti ETX atau penggunaan tenaga, yang mengarah kepada keseimbangan beban yang lebih baik dalam rangkaian. Untuk melaksanakan dan menilai metrik yang dicadangkan berbanding metrik RPL lain, ContikiOS dan COOJA simulator telah digunakan. Simulasi meluas telah dijalankan dengan cara yang sistematik yang menghasilkan analisis terperinci mengenai metrik yang diperkenalkan iaitu W-metrik, kiraan penghantaran dijangkakan (ETX) dan fungsi objektif sifar (OF0) yang menggunakan kiraan hop sebagai metrik penghalaan. Analisis dan perbandingan adalah  berdasarkan lima parameter prestasi, iaitu throughput, nisbah penghantaran paket (PDR), latency, panjang panjang antrian, dan penggunaan kuasa. Hasil simulasi menunjukkan bahawa W-metrik yang diperkenalkan mempunyai prestasi yang lebih baik berbanding dengan metrik RPL lain berkaitan dengan parameter prestasi yang dinyatakan di atas. Pada masa yang sama, hasil menunjukkan bahawa prestasi latency W-metrik adalah setanding dengan metrik penghalaan RPL yang lain. Dalam simulasi sampel 500 saat dengan 25 nod dan dengan nod yang menghantar paket secara berkala ke akar rangkaian pada kadar 1 paket setiap 4 saat, W-metrik menunjukkan keluaran yang sangat efisien iaitu 5.16 kbps, peningkatan sebanyak 8.2% berbanding ETX. Keputusan menunjukkan bahawa ia mempunyai nisbah penghantaran paket 93.3%, yang lebih tinggi berbanding 83.3% untuk ETX dan 74.2% untuk OF0. Purata panjang giliran 0.48 packet menunjukkan peningkatan 15.8% lebih baik daripada ETX. Di samping itu, ia mempamerkan penggunaan tenaga sebanyak 5.16 mW iaitu 2.1% kurang daripada ETX. Secara keseluruhan, W-metrik nampaknya menjadi alternatif yang berpotensi menggantikan ETX dan OF0 kerana ia memilih laluan yang lebih cekap dengan bekerja pada keseimbangan beban rangkaian dan dengan mempertimbangkan ciri-ciri pautan.

Research on Route Protocol and Architecture for WSN Based on IPv6

2012 ◽  
Vol 616-618 ◽  
pp. 2233-2238 ◽  
Author(s):  
Rui Ting Lu ◽  
Xiang Zhen Li ◽  
Jia Hui Wang ◽  
Feng Jie Sun
Keyword(s):  
Load Balancing ◽  
Routing Protocol ◽  
Network Performance ◽  
Experimental Result ◽  
Routing Overhead ◽  
On Demand ◽  
End To End Delay ◽  
Simulation Results ◽  
Demand Routing ◽  
Route Protocol

WSN based on IPv6 is a new network integrated IPv6 and WSN. The related technologies of WSN based on IPv6 was researched, and an architecture of WSN based on IPv6 was proposed according to 6LoWPAN protocol in this article. Efficient and stable route protocol is a main focus to ensure network performance. Refer to on-demand routing protocol DSR, a Load-Balancing route protocol for WSN based on IPv6 was designed. An implementation of this protocol was programmed in NS2, and its simulation results were analyzed. The experimental result shows that this protocol could effectively reduce end-to-end delay and routing overhead, improving the network performance.

Research of Load Unbalance of a Two- Motor Variable Frequency Dtive for Scraper Conveyor

2021 ◽  
Vol 64 (4-5) ◽  
pp. 37-45
Author(s):  
Dmitry Shprekher ◽  
◽  
Gennady Babokin ◽  
Evgeny Kolesnikov ◽  
Dmitry Ovsyannikov ◽  
...  
Keyword(s):  
Load Balancing ◽  
Electric Motor ◽  
Frequency Converter ◽  
Common Type ◽  
Single Frequency ◽  
Variable Frequency ◽  
Electric Motors ◽  
Electric Drives ◽  
Scraper Conveyor ◽  
Simulation Results

The article analyzes the uneven loading of the scraper conveyor electric motors. The most common type of multi-motor conveyor is considered here: two-drive, with head and end drives connected through gearboxes and sprockets by an endless chain with scrapers. The simulation results are presented for three variants of com-binations of parameters of electric drives. It has been established that the use of a two-motor variable frequen-cy drive of a scraper conveyor powered by a single frequency converter inevitably entails uneven loading of the electric motors of the head and end drives. To eliminate this disadvantage, it is proposed to supply each electric motor from its own frequency converter. A system of automatic load balancing is proposed. The implementation and the result of the work of the load balancing system are carried out on a model in the Matlab / Simulink system.

Sign in / Sign up

Export Citation Format

Share Document