A load balancing task allocation scheme in a hard real time system

Author(s):  
Jean Louis Lanet
Vestnik MEI ◽  
2018 ◽  
Vol 5 (5) ◽  
pp. 73-78
Author(s):  
Igor В. Fominykh ◽  
◽  
Sergey V. Romanchuk ◽  
Nikolay Р. Alekseev ◽  
◽  
...  

Author(s):  
Diego Pinheiro ◽  
Rawlinson Goncalves ◽  
Eduardo Valentin ◽  
Horacio Oliveira ◽  
Raimundo Barreto

2013 ◽  
Vol 475-476 ◽  
pp. 1161-1166
Author(s):  
Xu Dong Zhu ◽  
Hui You Chang

This paper puts forward a novel constraint specification. By limitations on the maximum of consecutive miss of deadline and the minimum probability of meeting deadline on the fixed sliding window s×m, the novel constraint specification has three contributions to the research of the weakly hard real-time system constraint specification. Firstly, has configurable flexible parameters which can easily realize all hard real-time, soft real-time and weakly hard real-time only by changing the parameters. Secondly, replacing the "fixed sliding window" with the "any" window, the novel constraint specification not only simplify the calculation of real-time, which will reduce the time consume, but also broaden the application scope such as smoothly scheduling. Moreover, the fixed sliding window is helpful to find a schedulable μ-pattern. Thirdly, the novel constraint specification solves the invalidity of some constraint specification when the first invocation of task is lost or missed. Through analysis and experiments, the results demonstrate that our novel constraint specification outperform previous approaches in both calculation methods and application scopes.


2014 ◽  
Vol 543-547 ◽  
pp. 1972-1976
Author(s):  
Huai Lin Dong ◽  
Ming Yuan He ◽  
Qing Feng Wu ◽  
Sheng Hang Wu

When membership queries are evaluated in a set, the performance can be improved by a Bloom filter which is a space-efficient probabilistic data structure. According to its space-efficient character, Bloom Filter presented to address the load balancing problem for streaming media information in Storm system which is free and open source distributed real time computation system. This method increases the server cluster availability by balancing the workloads among the servers within a cluster. Additionally, it improves real time system Storm efficiently in saving the data transmission time and reducing the calculation complexity.


Sign in / Sign up

Export Citation Format

Share Document