Distribuirano računanje pomoću programskog jezika Python

master's thesis

• Author: Martina Barišić
• Mentor: Robert Manger (mentor)
• Committee member: Robert Manger (predsjednik povjerenstva)
• Committee member: Goranka Nogo (član povjerenstva)
• Committee member: Goran Muić (član povjerenstva)
• Committee member: Filip Najman (član povjerenstva)
• Granter: University of Zagreb; Faculty of Science; (Department of Mathematics); Zagreb
• Defense date and country: 2018-09-24, Croatia
• Scientific / art field, discipline and subdiscipline: NATURAL SCIENCES; Mathematics

Abstract

U ovom radu proučavane su distribuirane aplikacije u programskom jeziku Python. Distribuirani, pa i paralelni sustavi, danas su široko primjenjivi. Prikazan je Amdahlov zakon koji govori da uložen napor u paralelizaciju postojećeg algoritma nadmašuje dobitke u performansama. Također su proučavani pojmovi višedretvenosti, višestrukih procesa i višeprocesorskih redova te neke njihove značajke i primjene. Pokazano je i kako se uspostavlja Celery pomoću kojeg je moguće razvijati distribuirane aplikacije i navedene su dvije njegove alternative. Na kraju rada, proučavane su dvije distribuirane aplikacije. Pokazalo se kako se vrši komunikacija između pojedinih dijelova aplikacija u određenim situacijama te što je potrebno promijeniti da rade na više računala. Praćeno je izvršavanje druge aplikacije (distribuirano sortiranje) te je zaključeno da, na velikom skupu podataka, bolje vrijeme izvršavanja ima distribuirana aplikacija sortiranja u odnosu na algoritam sortiranja koji se izvršavao na lokalnom računalu.

Abstract (english)

In this work, distributed applications are studied in the Python programming language. Distributed and parallel systems are today widely applied. Amdahl’s law is introduced, which states that the effort to parallelize the existing algorithm outweighs performance gains. Also, multiple threads, multiple processes and multiprocess queues are studied and some of their features and applications. It was also demonstrated how Celery is configured, by which it is possible to develop distributed applications and two of its alternatives are listed. At the end of the work, two distributed applications were studied. Communication between specific parts of the application in certain situations has been shown and what needs to be changed to work on multiple computers. The other application was executed (distributed sorting) and it was concluded that, on a large data set, better execution time is achieved by the distributed sorting application compared to the sorting algorithm executed on a local computer.

• Keywords: distribuirane aplikacije; Python; Amdahlov zakon; Celery; distribuirano sortiranje
• Keywords (english): distributed applications; Python; Amdahl’s law; Celery; distributed sorting
• Language: croatian
• URN:NBN: urn:nbn:hr:217:657008
• Study programme: Title: Computer Science and Mathematics; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra računarstva i matematike (magistar/magistra računarstva i matematike)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2019-03-04 10:35:21