Konkurējošu procesu veiktspējas uzlabošanas iespējas liela izmēra Oracle datubāzēs

Abstract

Darba apskatiti liela apjoma Oracle datubazes procesi, kuri darbojas gandriz reala laika (datiem sistema japaradas noteikta, isa laika posma pec notikuma reala dzive). Tika petita procesu veiktspeja un tas uzlabošanas iespejas, apskatita procesu savstarpeja mijiedarbiba un konkurence uz vieniem un tiem pašiem resursiem, piemeram, diska resursi, procesors, atmina. Darba aprakstits modelis, kurš prognoze procesa veiktspeju atkariba no procesa darama darba apjoma un sistemas resursiem. Tika apskatita operativas atminas datubažu uzbuve un izmantošanas iespejas veiktspejas uzlabošana. Darba aprakstiti veiktie veiktspejas petijumi, izmantojot Oracle datubazi un operativas atminas datubazi, ka ari sniegts ieskats veiktspejas prognozešanas modela izmantošana.

This work considers large volume Oracle database near real time processes (data in system appears in a very short time after event in real life). Author analyzed process’ performance and possibilities to improve it. Interaction of processes and competition for the same resource (HDD, processor, RAM) is inspected. Author describes model which forecast performance of process depending on work size and available system resources. In-memory databases and its uses in performance improvement are also analyzed in this work. Author describes performance investigations on Oracle database and In-memory database. In work is given insight into use of performance forecast model.