Prediction of cluster system load using adaptive model mixture

Y.S. Artamonov


In this paper, we examine the approach to choose a high-performance environment based on the prediction of the load of cluster nodes. The aim of the work is to investigate various prediction models in application to the task of forecasting the cluster load, choose the most successful model configurations and find out how to effectively apply these models all together.

The article presents the results of comparison of EMMSP models, models based on neural networks and adaptive models for solving the task of forecasting the load of cluster resources. The following parameters of neural network models are considered: selection of activation functions, algorithms for initialization and updating of neuron weights, and coding of additional features for training the network on the basis of date and time data. The testing of adaptive selection models and adaptive composition was performed and improvement of the forecasting results was shown in comparison with the models on which they were based. Training and testing of the models was performed using the load dataset for the cluster "Sergey Korolev" for the period from November 2013 to December 2016.

Full Text:

PDF (Russian)


S. Naseera, G.K. Rajini, P. Sunil Kumar Reddy “Host CPU Load Prediction Using Statistical Algorithms a comparative study” // International Journal of Computer Technology and Applications – 2016. – 9(12). – pp. 5577-5582.

S. Di, D. Kondo, W. Cirne “Host load prediction in a Google compute cloud with a Bayesian model” // Proceedings of the International Conference on High Performance Computing, Networking, Storage and Analysis. – IEEE Computer Society Press, 2012. – p. 21.

Ju.S. Artamonov “Primenenie modeli EMMSP dlja prognozirovanija dostupnyh vychislitel'nyh resursov v klasternyh sistemah” // Izvestija Samarskogo nauchnogo centra RAN. – 2016. – tom 18, # 4 (4). – S. 681-687.

S. Naseera, G.K. Rajini, N. Amutha Prabha, G. Abhishek “A comparative study on CPU load predictions in a computational grid using artificial neural network algorithms” // Indian Journal of Science and Technology. – 2015. – T. 8. – #. 35.

K. Kalaitzakis, G. Stavrakakis, E.M. Anagnostakis “Short-term load forecasting based on artificial neural networks parallel implementation” // Electric Power Systems Research. – 2002. – T. 63. – #. 3. – pp. 185-196.

M. Chandini, R. Pushpalatha, R. Boraia “A Brief study on Prediction of load in Cloud Environment” // International Journal of Advanced Research in Computer and Communication Engineering. – 2016. – 5(5). – pp. 157-162.

H.A. Engelbrecht, M. van Greunen “Forecasting methods for cloud hosted resources, a comparison” // Network and Service Management (CNSM), 2015 11th International Conference on. – IEEE, 2015. – pp. 29-35.

S. Hajkin Nejronnye seti. – M.: Vil'jams, 2006. – 1104 s.

Deeplearning4j: Open-source distributed deep learning for the JVM [Jelektronnyj resurs]. URL: (data obrashhenija: 01.01.2017)

C. Osovskij Nejronnye seti dlja obrabotki informacii. – M.: Finansy i statistika, 2002. – 344 s.

Y. Nesterov Introductory Lectures on Convex Optimization A Basic Course – Springer, 2004. – 211 p.

D.P. Kingma, J.L. Ba “ADAM: A Method for Stochastic Optimization” // arXiv: 1412:6980 [cs.LG], – 2014.

S. Ruder “An overview of gradient descent optimization algorithms” //arXiv preprint arXiv: 1609.04747. – 2016.

Ju.P. Lukashin Adaptivnye metody kratkosrochnogo prognozirovanija vremennyh rjadov. – M.: Finansy i statistika, 2003. – 415 s.

Ju.S. Artamonov, S.V. Vostokin “Razrabotka raspredelennyh prilozhenij sbora i analiza dannyh na baze mikroservisnoj arhitektury” // Izvestija Samarskogo nauchnogo centra Rossijskoj akademii nauk, t. 18, # 4(4), 2016. s.688-693.

Ju.S. Artamonov, S.V. Vostokin “Instrumental'noe programmnoe obespechenie dlja razrabotki i podderzhki ispolnenija prilozhenij nauchnyh vychislenij v klasternyh sistemah” // Vestn. Sam. gos. tehn. un-ta. Ser. Fiz.-mat. nauki, 19:4 (2015), S. 785–798.

O.S. Zaikin, M.A. Posypkin, A.A. Semjonov, N.P. Hrapov Opyt organizacii dobrovol'nyh vychislenij na primere proektov OPTIMA@home i SAT@home // Vestnik Nizhegorodskogo universiteta im. N.I. Lobachevskogo, 2012, 5(2), S. 340-347.

A. P. Afanasiev, I. V. Bychkov, M. O. Manzyuk, M. A. Posypkin, A. A. Semenov, O. S. Zaikin (2015). “Technology for integrating idle computing cluster resources into volunteer computing projects”. In Proc. of The 5th International Workshop on Computer Science and Engineering, Moscow, Russia (pp. 109-114).


  • There are currently no refbacks.

Abava  Absolutech Convergent 2020

ISSN: 2307-8162