Mathematical Methods for Optimization of Distributed Computing in a Heterogeneous Environment

Vladimir Novozhenov, Tatyana Romanova

Abstract


In this article, we examined the possible organization and architecture of a heterogeneous cluster environment with varying architectures, performance, and characteristics for distributed computing. Mathematical methods for optimizing task queue computation time in such an environment were also considered. The goal was to identify the optimal task scheduling algorithm for the cluster under given conditions and constraints of the mathematical model. For this purpose, were designed and applied a new method of task queue computing time optimization. As a result, the efficiency of the BlackFill algorithm, modified to meet the task requirements, was demonstrated. Numerical simulation results show the effectiveness of the proposed methods compared to traditional approaches. The application of the developed methods significantly reduces computation time, improves resource utilization, and ensures stable system performance under changing workloads. The results may be useful for the development of high-performance computing systems and management algorithms in heterogeneous environments

Full Text:

PDF (Russian)

References


V.A. Novozhenov, A.P. Kovtushenko. Creda obrabotki zadanij, posylaemyh na klaster iz vneshnego prilozhenija dlja uskorenija vychislenij, Nauchno-obrazovatel'nyj zhurnal dlja studentov i prepodavatelej «StudNet», #4, c. 244–251, 2020.

G.I. Radchenko. 2012. Raspredelennye vychislitel'nye sistemy.

Kwok Y.-K., Ahmad I. Static scheduling algorithms for allocating directed task graphs to multiprocessors. ACM Comput. Surv., vol. 31, # 4, pp. 406–471, 1999.

Singh K., Alam M., Sharma S.K. A survey of static scheduling algorithm for distributed computing system, Int. J. Comput. Appl. Foundation of Computer Science, vol. 129, # 2, pp. 25–30, 2015.

Alam M., Khan A., Varshney A.K. A review of dynamic scheduling algorithms for homogeneous and heterogeneous systems, Syst. Archit., pp. 73–83, 2018.

Chawla Y., Bhonsle M. A. Study on scheduling methods in cloud computing, Int. J. Emerg. Trends Technol. Comput. Sci., vol. 1, # 3, pp. 12–17, 2012.

Suhih A.V., Vasjaeva N.S. Issledovanie klassifikacij klasternyh sistem, Kibernetika i programmirovanie, # 2. c. 20–27, 2016.

Karpenko A.P. Mnogoprocessornye sistemy (MIMD-sistemy). Vychislitel'nye klastery. MGTU im. N.Je.Baumana [Jelektronnyj resurs] http://bigor.bmstu.ru/?cnt/?doc=Parallel/ch010104.mod/?cou=Parallel/base.cou

Salibekjan S.M., Aminev D.A., Semin V.G., 2012, Rezul'taty modelirovanija ob"ektno-atributnoj vychislitel'noj sistemy. Tezisy konferencii «Info-2012».

Salibekjan S.M., Panfilov P.B., Hakimullin E.R., Geterogennye vychislitel'nye sistemy na osnove ob"ektno-atributnoj modeli programmirovanija. https://www.hse.ru/pubs/share/direct/document/77867177

Suhoroslov O.V., Organizacija vychislenij v geterogennyh raspredelennyh sredah Izvestija JuFU. Tehnicheskie nauki. Razdel IV. Oblachnye vychislenija, s. 115–130.

Kondrat'ev A.A., Skrytye problemy raspredelennyh vychislenij, Electrical and data processing facilities and systems, # 3, t. 11, s. 61–65, 2015.

Chzho Za. 2020. Ispol'zovanie virtualizacii dlja uvelichenija jeffektivnosti vychislenii. Dissertacija na soiskanie uchenoj stepeni kandidata tehnicheskih nauk.

Grushin D.A., Kuzjurin N.N. O zadache jeffektivnogo upravlenija vychislitel'noj infrastrukturoj, Trudy ISP RAN, t. 30, # 6, s. 123–142, 2018.

Kuzjurin N.N., Grushin D.A., Fomin C.A. Problemy dvumernoj upakovki i zadachi optimizacii v raspredelennyh vychislitel'nyh sistemah. Trudy Instituta sistemnogo programmirovanija RAN. t. 26, # 1, s. 483–502, 2014.

Toporkov, V.V. 2004. Modeli raspredelennyh vychislenij. FIZMATLIT, 320 c.

Polezhaev P.N. Sovremennye algoritmy planirovanija zadach na klasternyh vychislitel'nyh sistemah. Metodologija i proektirovanie jekspertnyh sistem, sekcija # 2 «Aktual'nye problemy fundamental'noj i prikladnoj matematiki», s. 87–95.

Kovalenko V.N., Kovalenko E.I., Korjagin D.A., Semjachkin D.A., 2007. Upravlenie parallel'nymi zadanijami v gride s neotchuzhdaemymi resursami. Udalennyj resurs Instituta prikladnoj matematiki im. M.V. Keldysha Rossijskoj akademii nauk.

Sal'nikov, A.N. 2006. Sistema razrabotki i podderzhki ispolnenija parallel'nyh programm. Dissertacija na soiskanie uchenoj stepeni kandidata fiziko-matematicheskih nauk.

Mazalov V.V., Nikitina N.N. Ocenka harakteristik algoritma Backfill pri upravlenii potokom zadach na vychislitel'nom klastere. Vychislitel'nye tehnologii, t. 17, # 5, 2012.

Shevchenko S.V. Optimizacija raspredelennyh vychislenij v sistemah parallel'noj obrabotki dannyh, Intellektual'nye it v upravlenii, # 3, s. 25–28, 2018.

Kovalenko N.S. Model' organizacii vypolnenija sinhronnyh processov v mnogoprocessornyh sistemah i zadachi ih optimizacii. Matematicheskie struktury i modelirovanie t. 1, # 49, s. 80–88, 2019.

E.E. Perepelkin. 2013. Kompleksnoe modelirovanie i optimizacija uskoritel'nyh sistem na graficheskom processore (GPU). Dissertacija na soiskanie uchenoj stepeni doktora fiziko-matematicheskih nauk.

Makoshenko D.V. Matematichekaja model' vremeni vychislenij dlja optimizacii programm, Izvestija vuzov. Severo-Kavkazskij region. Estestvennye nauki, # 2, s. 10–13, 2009.

Pimenov A.V., Fedorov I.R., Bezzateev S.V. Postroenie arhitektury tumannyh vychislenij s ispol'zovaniem tehnologii blokchejn. Informacionno-upravljajushhie sistemy, # 5, s. 40–48, 2022.

Gorin D.I., Romanova T.N. Metod optimizacii koda dlja processora Qualcomm Hexagon, podderzhivajushhiĭ parallelizm na urovne komand i postroennyĭ po arhitekture VLIW (Very Long Instruction Word). Informacionnye tehnologii v nauke, obrazovanii i upravlenii, # 1, s. 105–116, 2021.


Refbacks

  • There are currently no refbacks.


Abava  Кибербезопасность IT Congress 2024

ISSN: 2307-8162