Performance Analysis of Wireless Sensor Networks [90öu6]
Project Lead
Project Duration
01/04/2015 - 31/03/2016Project URL
Go to WebsitePublications
2016
[Schreiner]
Analysis of Finite-Source Cluster Networks
Adam Toth, Tamas Berczes, Attila Kuki, Bela Almasi, Wolfgang Schreiner, Jinting Wang, Fang Wang
Creative Mathematics and Informatics 25(2), pp. 223-235. 2016. SINUS Association, ISSN 1584 - 286X.@article{RISC5217,
author = {Adam Toth and Tamas Berczes and Attila Kuki and Bela Almasi and Wolfgang Schreiner and Jinting Wang and Fang Wang},
title = {{Analysis of Finite-Source Cluster Networks}},
language = {english},
abstract = {Nowadays the distributed heterogeneous resources of networks, like the computational grid,start to have a greater part of interest so, the investigations of such systems are vital. Because of the more efficient utilisation of the resources, the job scheduling becomes more challenging for the system administrators. The allocation of the arriving jobs has a great impact on the efficiency and the energy consumption of the system. In this paper, we present a finite source generalized model for the performance evaluation of scheduling compute-intensive jobs based on the infinite model of Tien Van Do. The available computers are classified into three groups. This classification is based on two aspects: high performance priority (HP) and energy efficiency priority (EE). We investigate three schemes (separate queue, class queue and common queue) for buffering the jobs in a computational cluster that is built from Commercial Off-The-Shelf (COTS) servers. Our main interest is to calculate performance measures and energy consumption of the system using the different buffering schemes and classifications.},
journal = {Creative Mathematics and Informatics},
volume = {25},
number = {2},
pages = {223--235},
publisher = {SINUS Association},
isbn_issn = {ISSN 1584 - 286X},
year = {2016},
refereed = {yes},
sponsor = {Stiftung Aktion Österreich-Ungarn contract 90öu6. },
length = {14}
}
author = {Adam Toth and Tamas Berczes and Attila Kuki and Bela Almasi and Wolfgang Schreiner and Jinting Wang and Fang Wang},
title = {{Analysis of Finite-Source Cluster Networks}},
language = {english},
abstract = {Nowadays the distributed heterogeneous resources of networks, like the computational grid,start to have a greater part of interest so, the investigations of such systems are vital. Because of the more efficient utilisation of the resources, the job scheduling becomes more challenging for the system administrators. The allocation of the arriving jobs has a great impact on the efficiency and the energy consumption of the system. In this paper, we present a finite source generalized model for the performance evaluation of scheduling compute-intensive jobs based on the infinite model of Tien Van Do. The available computers are classified into three groups. This classification is based on two aspects: high performance priority (HP) and energy efficiency priority (EE). We investigate three schemes (separate queue, class queue and common queue) for buffering the jobs in a computational cluster that is built from Commercial Off-The-Shelf (COTS) servers. Our main interest is to calculate performance measures and energy consumption of the system using the different buffering schemes and classifications.},
journal = {Creative Mathematics and Informatics},
volume = {25},
number = {2},
pages = {223--235},
publisher = {SINUS Association},
isbn_issn = {ISSN 1584 - 286X},
year = {2016},
refereed = {yes},
sponsor = {Stiftung Aktion Österreich-Ungarn contract 90öu6. },
length = {14}
}
2015
[Schreiner]
Modeling RF Communication in Sensor Networks by Probabilistic Model Checking
Wolfgang Schreiner, Tamas Berczes, Janos Sztrik, Adam Toth
Research Institute for Symbolic Computation (RISC), Johannes Kepler University, Linz, Austria. Technical report no. 15-21, October 2015. [pdf]@techreport{RISC5176,
author = {Wolfgang Schreiner and Tamas Berczes and Janos Sztrik and Adam Toth},
title = {{Modeling RF Communication in Sensor Networks by Probabilistic Model Checking}},
language = {english},
abstract = {We report in this paper our results of modeling and analyzing with the probabilistic modelchecker PRISM a system of radio frequency (RF) transmission in sensor networks whichhas previously been studied in literature by using finite-source retrial queueing systems.We are able to validate with a small and quite transparent PRISM model the previouslyreported results (and also exhibit a minor error). Furthermore, we extend the model by alsoconsidering infinite sources and show that a previously suggested optimization has in thismodel beneficial effects only in a comparatively small parameter range.},
number = {15-21},
year = {2015},
month = {October},
institution = {Research Institute for Symbolic Computation (RISC), Johannes Kepler University, Linz, Austria},
sponsor = {Supported by the project 90öu6 “Leistungsmodellierung von Drahtlosen Sensor-Netzwerken” of the Stiftung Ak- tion Österreich-Ungarn},
length = {25}
}
author = {Wolfgang Schreiner and Tamas Berczes and Janos Sztrik and Adam Toth},
title = {{Modeling RF Communication in Sensor Networks by Probabilistic Model Checking}},
language = {english},
abstract = {We report in this paper our results of modeling and analyzing with the probabilistic modelchecker PRISM a system of radio frequency (RF) transmission in sensor networks whichhas previously been studied in literature by using finite-source retrial queueing systems.We are able to validate with a small and quite transparent PRISM model the previouslyreported results (and also exhibit a minor error). Furthermore, we extend the model by alsoconsidering infinite sources and show that a previously suggested optimization has in thismodel beneficial effects only in a comparatively small parameter range.},
number = {15-21},
year = {2015},
month = {October},
institution = {Research Institute for Symbolic Computation (RISC), Johannes Kepler University, Linz, Austria},
sponsor = {Supported by the project 90öu6 “Leistungsmodellierung von Drahtlosen Sensor-Netzwerken” of the Stiftung Ak- tion Österreich-Ungarn},
length = {25}
}
[Schreiner]
Analyzing Cluster Scheduling Schemes by Probabilistic Model Checking (Addendum)
Wolfgang Schreiner, Tamas Berczes, Janos Sztrik, Adam Toth
Research Institute for Symbolic Computation (RISC), Johannes Kepler University, Linz, Austria. Technical report, October 2015. [pdf]@techreport{RISC5177,
author = {Wolfgang Schreiner and Tamas Berczes and Janos Sztrik and Adam Toth},
title = {{Analyzing Cluster Scheduling Schemes by Probabilistic Model Checking (Addendum)}},
language = {english},
abstract = {This short note presents an addendum to the previously published report “Analyzing Clus-ter Scheduling Schemes by Probabilistic Model Checking” in which we have used the prob-abilistic model checker PRISM various cluster scheduling schemes that were introduced byDo, Vu, Tran, and Nguyen in their paper “A generalized model for investigating schedulingschemes in computational clusters” and analyzed by simulation there. Here we model andanalyze two new policies for the “Separate Queue” scheme where each server has a separatejob queue; in these policies the mean response time of new jobs is taken into account. Itis shown that the policies may perform better than the original one; for a certain range ofparameters, however, they also may perform worse.},
year = {2015},
month = {October},
institution = {Research Institute for Symbolic Computation (RISC), Johannes Kepler University, Linz, Austria},
sponsor = {Supported by the project 90öu6 “Leistungsmodellierung von Drahtlosen Sensor-Netzwerken” of the Stiftung Aktion Österreich-Ungarn},
length = {17}
}
author = {Wolfgang Schreiner and Tamas Berczes and Janos Sztrik and Adam Toth},
title = {{Analyzing Cluster Scheduling Schemes by Probabilistic Model Checking (Addendum)}},
language = {english},
abstract = {This short note presents an addendum to the previously published report “Analyzing Clus-ter Scheduling Schemes by Probabilistic Model Checking” in which we have used the prob-abilistic model checker PRISM various cluster scheduling schemes that were introduced byDo, Vu, Tran, and Nguyen in their paper “A generalized model for investigating schedulingschemes in computational clusters” and analyzed by simulation there. Here we model andanalyze two new policies for the “Separate Queue” scheme where each server has a separatejob queue; in these policies the mean response time of new jobs is taken into account. Itis shown that the policies may perform better than the original one; for a certain range ofparameters, however, they also may perform worse.},
year = {2015},
month = {October},
institution = {Research Institute for Symbolic Computation (RISC), Johannes Kepler University, Linz, Austria},
sponsor = {Supported by the project 90öu6 “Leistungsmodellierung von Drahtlosen Sensor-Netzwerken” of the Stiftung Aktion Österreich-Ungarn},
length = {17}
}
