A.P. Gantimurov – Post-graduate Student, 

Bauman Moscow State Technical University

E-mail: a.gantimurov@npobaum.ru

A.D. Vinichenko – Part-programming Engineer, 

LLC «SK-SHD» (Moscow)

E-mail: a.vinichenko@sk-shd.ru

I.P. Ivanov – Dr.Sc.(Eng.), Associate Professor, 

Bauman Moscow State Technical University

E-mail: ivanov@bmstu.ru

A.V. Bosov – Engineer, 

IEC «Composites of Russia», Bauman Moscow State Technical University E-mail: aleksey.bosov@emtc.ru

A mandatory component of modern data storage and processing systems is the self-diagnosis subsystem, the functions of which include identifying malfunctions in the system components and timely notification of personnel performing system administration. Due to the specifics of a distributed data storage system, the distribution of stored information is carried out evenly throughout the entire available array of hard drives.

The task of detecting and preventing failures in today’s distributed storage systems is an important task in its operation and maintenance. Method of early diagnostics of disk failure in distributed data storage system using mathematical apparatus of fuzzy systems theory is proposed. Analysis of dynamics of data exchange buffers filling between Control module and Storage modules and application of mathematical apparatus of fuzzy systems theory was performed. Results of mathematical simulation of situations of defective disk appearance in hard disk array during recording of stored information into distributed data storage system in real time are given. The early diagnosis method allows you to initiate mechanisms in advance to minimize the overall performance loss of a distributed storage system.

1. Avizhenis A. Otkazoustoychivost’ – svoystvo, obespechivayushchee postoyannuyu rabotosposobnost’ tsifrovykh sistem. TIIER. 1978. T. 66. № 10. S. 5−25. (In Russian).
2. Vedeshenkov V.A. Samodiagnostirovanie tsifrovykh sistem s rekonfiguratsiey. Problemy upravleniya. № 4. S. 39−51. (In Russian).
3. Fedorov A.R. Sposoby kodirovaniya informatsii dlya postroeniya programmnykh otkazoustoychivykh diskovykh massivov. Sistemnoe rogrammirovanie. 2012. T. 7. S. 6−31. (In Russian).
4. Sage A. Weil, Scott A. Brandt, Ethan L. Miller, Carlos Maltzahn. CRUSH: Controlled, Scalable, Decentralized Placement of Replicated Data [Elektronnyy resurs]. URL = https://ceph.io/wp-content/uploads/2016/08/weil-crush-sc06.pdf.
5. Kruglov V.V., Dli M.I. Intellektual’nye informatsionnye sistemy: komp’yuternaya podderzhka sistem nechetkoy logiki i nechetkogo vyvoda. M.: Fizmatlit. 2002. (In Russian).
6. Leolenkov A.V. Nechetkoe modelirovanie v srede MATLAB i fuzzyTECH. SPb. 2003. (In Russian).
7. Rutkovskaya D., Pilin’skiy M., Rutkovskiy L. Neyronnye seti, geneticheskie algoritmy i nechetkie sistemy. M. 2004. (In Russian).
8. Ivanov I.P., Gantimurov A.P., Vinichenko A.D., Bosov A.V. Vybor arkhitekturnykh resheniy postroeniya sistemy khraneniya dannykh v tekhnologii Web-Scale IT. Perspektivy nauki. 2019. № 8(119). S. 57−62. (In Russian).