Cluster system load balancing model with consideration of hardware characteristics of server hardware
https://doi.org/10.37661/1816-0301-2022-19-4-84-93
Abstract
Objectives. To upgrade and complement the existing load balancing model in multi-server systems, taking into account the hardware characteristics of the server equipment, as well as the most loaded components of the server equipment in the video conferencing service cluster in educational processes and distance education.
Methods. The existing mathematical model of load balancing as a mass exchange system is considered, when significant changes are introduced: penalties for equipment downtime and penalties for waiting in a queue will depend on the load on the server hardware components in the cluster architecture of video conferencing service.
Results. Formulas are given for calculating the total performance of a cluster of n servers with the maximum and minimum load of server hardware components in a videoconferencing system cluster.
Conclusion. A modeling complex has been developed to test the mathematical model on a system of up to n < 10 servers in a cluster of a videoconferencing system. Based on the results of calculations of the modeling complex, it was concluded that it is necessary to upgrate the existing algorithm for balancing the load on the selected BigBlueButton video conferencing service.
About the Author
A. N. Markov
Center for Informatization and Innovation Development of the Belarusian State University of Informatics and Radioelectronics
Belarus
Belarus
Aleksey N. Markov, Postgraduate Student, Deputy Head of the Center for Informatization and Innovative Development
st. Platonova, 39, Minsk, 220013
References
1. Paramonov A. I. Distance education problems and their applied solutions in educational technologies. Vysshee tehnicheskoe obrazovanie: problemy i puti razvitija : materialy Х Mezhdunarodnoj nauchnometodicheskoj konferencii, Minsk, 26 nojabrja 2020 g. [Engineering Education: Challendes and Developments : Materials of the X International Scientific and Methodological Conference, Minsk, 26 November 2020]. Minsk, Belorusskij gosudarstvennyj universitet informatiki i radiojelektroniki, 2020, рр. 182–187 (In Russ.).
2. Markov A. N., Migalevich S. A. Readiness of higher education institutions for digital transformation processes. Cifrovaja transformacija [Digital Transformation], 2021, no. 2, рр. 64–68 (In Russ.).
3. Markov A. N., Ihnatovich R. O., Paramonov A. I. Choosing a video conferencing service and its adaptation for educational institution. Informatika [Informatics], 2021, vol. 18, no. 4, рр. 17–25 (In Russ.). https://doi.org/10.37661/1816-0301-2021-18-4-17-25
4. Lui C., Albitz P. DNS and BIND, 5th edition. O'Reilly Media, Inc., 2006, 642 р.
5. Starykh V., Varyonov D., Ploskov S., Kuznetsov A. Construction and research of a distributed information system of web applications and services based on cluster architecture. Kachestvo. Innovacii. Obrazovanie [Quality. Innovation. Education], 2011, no. 12, рр. 91–99 (In Russ.).
6. Owens J. D., Luebke D., Govindaraju N., Harris M., Kr˝uger J., …, Purcell T. J. A survey of generalpurpose computation on graphics hardware. Computer Graphics Forum, 2007, vol. 26, no. 1, рр. 80–113.
7. Dongarra J. J., Luszczek P., Petitet A. The LINPACK benchmark: Past, present and future. Concurrency and Computation: Practice and Experience, 2003, vol. 15, no. 9, рр. 803–820.
8. Dongarra J., Grandinetti L., Joubert G. R., Kowalik J. High Performance Computing: Technology, Methods and Applications. Inc., 2015, 444 p.
9. Dongarra J. Adaptive linear solvers and eigensolvers. Argonne Training Program on Extreme-Scale Computing, 2016. Available at: https://extremecomputingtraining.anl.gov/files/2016/08/Dongara_830_ AdaptiveLinear.pdf (аccessed 08.08.2022).
Review
For citations:
Markov A.N. Cluster system load balancing model with consideration of hardware characteristics of server hardware. Informatics. 2022;19(4):84-93. (In Russ.) https://doi.org/10.37661/1816-0301-2022-19-4-84-93
Views: 241
Email this article 