Primjena baza podataka zasnovanih na vremenskim sljedovima u Internetu stvari

master's thesis

• Title (english): Time Series Databases Application in the Internet of Things
• Author: Mario Slatinac
• Mentor: Mario Kušek (mentor)
• Committee member: Mario Kušek (predsjednik povjerenstva)
• Committee member: Jurica Babić (član povjerenstva)
• Committee member: Krešimir Pripužić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2019-07-09, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Primjena baza podataka zasnovanih na vremenskim sljedovima u Internetu stvari sve je popularnija zbog specijalizirane primjene takvih baza podataka na vremensko slijedne podatke koji dolaze iz Interneta stvari. Među vremensko slijednim bazama podataka istaknule su se baze podataka InfluxDB i TimescaleDB koja je proširenje na bazu podataka PostgreSQL. Prednost vremensko slijednih baza podataka nad NoSQL i relacijskim bazama podataka je u visokim performancama, funkcijama namijenjenim za obradu i upravljanje vremensko slijednim podacima kao što su kontinuirani upiti koji se periodično i automatski izvode nad podacima koji pristižu u stvarnom vremenu i čiji se rezultat pohranjuje u bazu podataka. Korištenjem alata Time Series Benchmark Suite provedeno je testiranje performanci unosa podataka s obzirom na veličinu grupe podataka i kardinalnost skupa podataka za bazu podataka InfluxDB i TimescaleDB. Rezultati testiranja pokazali su da baza podataka InfluxDB ima manje vrijeme izvođenja od baze podatka TimescaleDB s povećanjem veličine grupe. Oba dvije baze podataka pokazuju povećanje vremena unosa podataka s povećanjem kardinalnosti skupa podataka. Baza podataka InfluxDB u oba dva testna scenarija pokazuje visoku potrošnju računalnih resursa, odnosno procesora i radne memorije dok kod baze podataka TimescaleDB to nije slučaj.

Abstract (english)

The application of time series databases on the Internet of things is increasingly popular because of the specialized use of such databases to time series data coming from the Internet of things.
Among the time series databases, InfuxDB and TimescaleDB databases have pointed out, where TimescaleDB is an extension for the PostgreSQL database.
The advantage of time series databases over NoSQL and relational databases is in high performance, functions designed for aggregating and managing time series data such as continuous queries that are periodically and automatically executed over real-time data and whose result is stored in the database.
By using the Time Series Benchmark Suite tool, performance of data ingest were tested with different batch size and different cardinality of data set for InfluxDB and TimescaleDB database.
Test results have shown that the InfluxDB database needs less time to ingest data than the TimescaleDB database when batch size is increasing.
Both databases show increased data ingest time by growing cardinality for the dataset. The InfluxDB database in both test scenarios shows the high usage of CPU and random-access memory (RAM), where for the TimescaleDB it is not the case.

• Keywords: baza podatka; vremensko slijedne baze podataka; InfluxDB; TimescaleDB; Telegraf; Grafana; kontinuirani upiti
• Keywords (english): database; time series databases; time-series; InfluxDB; TimescaleDB; Telegraf; Grafana; continuous query
• Language: croatian
• URN:NBN: urn:nbn:hr:168:694017
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2020-01-07 22:03:14