Raspodijeljeni geoprostorni sustav objavi-pretplati s visokom učestalošću objavljivanja

doctoral thesis

• Title (english): Distributed geospatial publish/subscribe system with high publication frequency
• Author: Ivan Livaja
• Mentor: Krešimir Pripužić (mentor)
• Committee member: Dejan Škvorc (predsjednik povjerenstva)
• Committee member: Martina Antonić (član povjerenstva)
• Committee member: Goran Martinović (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; (Department of Telecommunications); Zagreb
• Defense date and country: 2023-10-23, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing
• Universal decimal classification (UDC): 004 - Computer science and technology. Computing. Data processing

Abstract

Sustav objavi-pretplati je raspodijeljeni komunikacijski sustav u kojem se proizvođači poruka nazivaju objavljivači te objavljuju poruke (objave) koje žele isporučiti potrošačima poruka, koji se nazivaju pretplatnicima. Pretplatnici se moraju unaprijed pretplatiti na poruke od interesa kako bi ih mogli primiti nakon objavljivanja. Ovaj doktorski rad se bavi posebnom vrstom sustava objavi-pretplati, a to su geoprostorni sustavi objavi-pretplati u kojima i objave i pretplate sadrže geoprostorni objekt. Takav objekt se u objavama koristi za izražavanje informacija o njihovoj lokaciji, a u pretplatama za izražavanje lokacije interesa. Postoji veliki potencijal za korištenje geoprostornih sustava objavi-pretplati u Internetu stvari te na senzorskom Webu. Međutim postojeći geoprostorni sustavi objavi-pretplati nisu primjenjivi u tu svrhu jer su centralizirani i ne mogu se nositi s tokovima geoprostornih podataka koji imaju visoku učestalost objavljivanja. Kako bi se prevladalo ovo ograničenje, u doktorskom radu se predstavlja raspodijeljeni geoprostorni sustav objavi-pretplati za grozd računala koji u stvarnom vremenu učinkovito uspoređuje dolazne objave sa skupom pohranjenih pretplata. Osim toga, predložena su četiri različita algoritma uparivanja objava i pretplata u raspodijeljenom geoprostornom sustavu objavi-pretplati. Također su predstavljeni rezultati opsežne eksperimentalne studije u kojoj se uspoređuje propusnost, kašnjenje i potrošnja memorije ovih algoritama. Ovi rezultati jasno pokazuju da su predloženi algoritmi učinkoviti i skalabilni na veće grozdove. Usporedba s najsuvremenijim centraliziranim pristupima pokazuje da su dodatni troškovi obrade predloženih raspodijeljenih algoritama koje uvodi platforma Apache Spark gotovo zanemarivi.

Abstract (english)

Publish/subscribe is a messaging pattern where message producers, called publishers, publish messages which they want to be distributed to message consumers, called subscribers. Subscribers are required to subscribe to messages of interest in advance to be able to receive them upon publishing. In this paper, we discuss a special type of publish/subscribe systems, namely geospatial publish/subscribe systems, in which both published messages (i.e., publications) and subscriptions include a geospatial object. Such an object is used to express both the location information of a publication and the location of interest of a subscription. We argue that there is a great potential for using geospatial publish/subscribe systems for the Internet of Things and Sensor Web applications. However, existing geospatial publish/subscribe systems are not applicable for this purpose since they are centralized and cannot cope with multiple highly frequent incoming publications containing geospatial data streams. To overcome this limitation, we present a distributed geospatial publish/subscribe system in the cluster which efficiently matches incoming publications in real-time with a set of stored subscriptions. Additionally, we propose four different (distributed) algorithms for matchinging publications and subscriptions in our distributed geospatial publish/subscribe system. Finally, we present results of an extensive experimental evaluation in which we compare the throughput, latency, and memory consumption of these algorithms. These results clearly show that they are both efficient and scalable to larger clusters. The comparison with centralized state-of-the-art approaches shows that the processing overhead of our distributed algorithms introduced by the Apache Spark platform is almost negligible.

• Keywords: geoprostorni podaci; particioniranje; replikacija podataka; veliki skupovi podataka; obrada tokova podataka
• Keywords (english): geospatial data; partitioning; data replication; big data; data stream processing
• Language: croatian
• URN:NBN: urn:nbn:hr:168:063909
• Promotion: 2023
• Study programme: Title: Computer Science; Study programme type: university; Study level: postgraduate; Academic / professional title: Doktor znanosti (Doktor znanosti)
• Type of resource: Text
• Extent: 116 str. : graf. prikazi u boji ; 30 cm
• File origin: Born digital
• Access conditions: Open access
• Created on: 2024-05-10 13:33:11