prikaz prve stranice dokumenta Integracija i testiranje skalabilne prostorno-vremenske platforme Geomesa na Kubernetes platformi
Access restricted to authenticated users
master's thesis
Integracija i testiranje skalabilne prostorno-vremenske platforme Geomesa na Kubernetes platformi
2021. urn:nbn:hr:179:569668
Židić, Dinko
University of Split
Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture
Department of Electronics and Computing

Cite this document

Židić, D. (2021). Integracija i testiranje skalabilne prostorno-vremenske platforme Geomesa na Kubernetes platformi (Master's thesis). Retrieved from https://urn.nsk.hr/urn:nbn:hr:179:569668

Židić, Dinko. "Integracija i testiranje skalabilne prostorno-vremenske platforme Geomesa na Kubernetes platformi." Master's thesis, University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, 2021. https://urn.nsk.hr/urn:nbn:hr:179:569668

Židić, Dinko. "Integracija i testiranje skalabilne prostorno-vremenske platforme Geomesa na Kubernetes platformi." Master's thesis, University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, 2021. https://urn.nsk.hr/urn:nbn:hr:179:569668

Židić, D. (2021). 'Integracija i testiranje skalabilne prostorno-vremenske platforme Geomesa na Kubernetes platformi', Master's thesis, University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, accessed 21 March 2023, https://urn.nsk.hr/urn:nbn:hr:179:569668

Židić D. Integracija i testiranje skalabilne prostorno-vremenske platforme Geomesa na Kubernetes platformi [Master's thesis]. Split: University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture; 2021 [cited 2023 March 21] Available at: https://urn.nsk.hr/urn:nbn:hr:179:569668

D. Židić, "Integracija i testiranje skalabilne prostorno-vremenske platforme Geomesa na Kubernetes platformi", Master's thesis, University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, Split, 2021. Available at: https://urn.nsk.hr/urn:nbn:hr:179:569668

Please login to the repository to save this object to your list.
Metadata
TitleIntegracija i testiranje skalabilne prostorno-vremenske platforme Geomesa na Kubernetes platformi
Title (english)Building and benchmarking a scalable spatio-temporal analytics platform Geomesa on Kubernetes
AuthorDinko Židić
MentorMladen Russo (mentor)
GranterUniversity of Split
Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture
(Department of Electronics and Computing)
Split
Defense date and country2021-09-22, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Computing
Abstract
Tema diplomskog rada je podizanje i testiranje prostorno-vremenske platforme u obliku mikroservisne arhitekture koja se sastoji od Spark-a, Accumula s Geomesom koja predstavlja prostorno-vremensku bazu podataka čije su performanse testirane, a koristi Zookeeper i Hadoop za koordinaciju i pohranu podataka. Prva platforma za razvoj mikroservisne arhitekture u obliku kontejnera je Docker. Potreba za orkestracijom kontejnera dovela je do razvoja Kubernetes-a na kojem je podignuta cijela platforma. Kako bi se ispitale Geomesa performanse provedeni su vremenski, prostorni i prostorno-vremenski upiti u obliku dohvata i prebrojavanja podataka. Svi su izvršeni kako bi se provelo prostorno, vremensko, prostorno-vremensko testiranje i testiranje skaliranja koje uključuje skaliranje broja dohvaćenih podataka i dodijeljenih resursa. Rezultati su pokazali da Geomesa ima različite performanse za različite tipove upita. Najbolje se pokazala za prostorne, dok je za vremenske upite imala nestabilne odazive. Vrijeme odaziva je bilo manje od minute iako je broj podataka veći od dva milijuna. Pokazano je da bolje podnosi upite kod kojih su prostorni uvjeti velike površine, te da je odaziv ovisan o prostornoj distribuciji podataka. Uzrok tome je indeksiranje podatka po prostoru umjesto po podatcima. Pokazana je i važnost korištenja gigabitne mreže, te da nema poboljšanja odaziva kod korištenja SSD-a.
Abstract (english)
Master thesis in this work is building scalable spatio-temporal platform in microservice form, which consists of Spark, Accumulo with Geomesa spatio-temporal database whose, performance was examined. Accumulo requires Zookeeper and Hadoop for distributed coordination and data storage. First platform for developing applications in container in a form of microservice is Docker. Due to need, for container orchestration, Kubernetes on which whole platform is deployed, is developed. To examine Geomesa performance, spatial, temporal and spatio-temporal queries are performed as data count and fetch. All queries are performed to evaluate spatial, temporal, spatio-temporal benchmarks, and scalability benchmark that includes scaling fetched data volume and scaling available resources. Results show that Geomesa response time has high dependency on a type of query. Best response time is for spatial, while temporal benchmark response time is unstable. Response time for all benchmarks is under a minute, thus data volume is over two million records. Geomesa gives better response times for queries with larger area, and response time is depending on spatial data distribution. Reason for that is Geomesa indexing schema, which splits indexes with regards to space rather than data, resulting with unevenly balanced buckets. Results also show importance of using gigabit network, and no better response time with SSD disk.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:179:569668
Study programmeTitle: Computing
Study programme type: university
Study level: graduate
Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
Type of resourceText
File originBorn digital
Access conditionsAccess restricted to authenticated users
Terms of useLicence In copyright icon
Public noteRad izrađen u suradnji s Ericsson Nikola Tesla d.d.
Created on2022-07-06 08:27:59