Brīvu mākoņresursu pieejamība datu analīzei

Abstract

Mākoņresursi ir ļoti izplatīti un paliek ar vien populārāki un uztur ļoti lielu daļu internetā pieejamo resursu, kā rezultāta tiek veidotas lielas infrastruktūras šīs informācijas nodrošināšanai. Liela daļa resursu no šīm infrastruktūrām netiek pilnībā izmantoti. Tas liecina, ka ir daudz brīvi pieejamu resursu, kurus var potenciāli izmantot datu analīzei. Tādēļ bakalaura darba izstrādes gaitā tiks meklēts un izstrādāts risinājums, ar kuru palīdzību veicot liela apjoma datu apstrādi paralēli citiem procesiem, tie neietekmē galveno procesu veiktspēju. Risinājums tiek balstīts uz to, ka brīvu mākoņresursu pieejamais daudzums ir aptuveni 70% un tie ir potenciāli izmantojami datu analīzei, veidojot risinājumu uz VMware vSphere bāzes, kopā ar Kubernetes un Docker konteineriem, veiksmīgi izmantojot līdz 17% CPU jaudas nepārsniedzot 5% CPU gaidīšanas laika.

Availability of free cloud resources for data analysis Cloud resources are very common and they are becoming more and more popular, and they maintain a very large share of the resources available on the Internet, resulting in the creation of large infrastructures that can provide and store the information found on Internet. Much of the resources from these infrastructures are underused. This suggests that there are many freely available resources that can potentially be used for data analysis. Therfore, during the development of the bachelor’s thesis, a solution will be sought and developed, that will help with launching large-scale data processing processes in parallel with other processes. Solution is based on that, that available unused cloud resourses are around 70% and that they are potentially usable for data analytics, and it’s made so that data analytics do not affect the performance of the main key processes. This solution was created based on VMware vSphere, together with Kubernetes and Docker containers, to successfully use 20% additional CPU resources without reaching CPU readiness above 5%.