Process management for Internet of mobile things
Failid
Kuupäev
2022-11-08
Autorid
Ajakirja pealkiri
Ajakirja ISSN
Köite pealkiri
Kirjastaja
Abstrakt
Värkvõrk, ehk Asjade Internet (Internet of Things, lüh IoT) edendab lahendusi nagu nn tark linn, kus meid igapäevaselt ümbritsevad objektid on ühendatud infosüsteemidega ja ka üksteisega. Selliseks näiteks võib olla teekatete seisukorra monitoorimissüsteem. Võrku ühendatud sõidukitelt (nt bussidelt) kogutakse videomaterjali, mida seejärel töödeldakse, et tuvastada löökauke või lume kogunemist. Tavaliselt hõlmab selline lahendus keeruka tsentraalse süsteemi ehitamist. Otsuste langetamiseks (nt milliseid sõidukeid parasjagu protsessi kaasata) vajab keskne süsteem pidevat ühendust kõigi IoT seadmetega. Seadmete hulga kasvades võib keskne lahendus aga muutuda pudelikaelaks.
Selliste protsesside disaini, haldust, automatiseerimist ja seiret hõlbustavad märkimisväärselt äriprotsesside halduse (Business Process Management, lüh BPM) valdkonna standardid ja tööriistad. Paraku ei ole BPM tehnoloogiad koheselt kasutatavad uute paradigmadega nagu Udu- ja Servaarvutus, mis tuleviku värkvõrgu jaoks vajalikud on. Nende puhul liigub suur osa otsustustest ja arvutustest üksikutest andmekeskustest servavõrgu seadmetele, mis asuvad lõppkasutajatele ja IoT seadmetele lähemal. Videotöötlust võiks teostada mini-andmekeskustes, mis on paigaldatud üle linna, näiteks bussipeatustesse.
Arvestades IoT seadmete üha suurenevat hulka, vähendab selline koormuse jaotamine vähendab riski, et tsentraalne andmekeskust ülekoormamist.
Doktoritöö uurib, kuidas mobiilsusega seonduvaid IoT protsesse taoliselt ümber korraldada, kohanedes pidevalt muutlikule, liikuvate seadmetega täidetud servavõrgule. Nimelt on ühendused katkendlikud, mistõttu otsuste langetus ja planeerimine peavad arvestama muuhulgas mobiilseadmete liikumistrajektoore. Töö raames valminud prototüüpe testiti Android seadmetel ja simulatsioonides. Lisaks valmis tööriistakomplekt STEP-ONE, mis võimaldab teadlastel hõlpsalt simuleerida ja analüüsida taolisi probleeme erinevais realistlikes stsenaariumites nagu seda on tark linn.
The Internet of Things (IoT) promotes solutions such as a smart city, where everyday objects connect with info systems and each other. One example is a road condition monitoring system, where connected vehicles, such as buses, capture video, which is then processed to detect potholes and snow build-up. Building such a solution typically involves establishing a complex centralised system. The centralised approach may become a bottleneck as the number of IoT devices keeps growing. It relies on constant connectivity to all involved devices to make decisions, such as which vehicles to involve in the process. Designing, automating, managing, and monitoring such processes can greatly be supported using the standards and software systems provided by the field of Business Process Management (BPM). However, BPM techniques are not directly applicable to new computing paradigms, such as Fog Computing and Edge Computing, on which the future of IoT relies. Here, a lot of decision-making and processing is moved from central data-centers to devices in the network edge, near the end-users and IoT sensors. For example, video could be processed in mini-datacenters deployed throughout the city, e.g., at bus stops. This load distribution reduces the risk of the ever-growing number of IoT devices overloading the data center. This thesis studies how to reorganise the process execution in this decentralised fashion, where processes must dynamically adapt to the volatile edge environment filled with moving devices. Namely, connectivity is intermittent, so decision-making and planning need to involve factors such as the movement trajectories of mobile devices. We examined this issue in simulations and with a prototype for Android smartphones. We also showcase the STEP-ONE toolset, allowing researchers to conveniently simulate and analyse these issues in different realistic scenarios, such as those in a smart city.
The Internet of Things (IoT) promotes solutions such as a smart city, where everyday objects connect with info systems and each other. One example is a road condition monitoring system, where connected vehicles, such as buses, capture video, which is then processed to detect potholes and snow build-up. Building such a solution typically involves establishing a complex centralised system. The centralised approach may become a bottleneck as the number of IoT devices keeps growing. It relies on constant connectivity to all involved devices to make decisions, such as which vehicles to involve in the process. Designing, automating, managing, and monitoring such processes can greatly be supported using the standards and software systems provided by the field of Business Process Management (BPM). However, BPM techniques are not directly applicable to new computing paradigms, such as Fog Computing and Edge Computing, on which the future of IoT relies. Here, a lot of decision-making and processing is moved from central data-centers to devices in the network edge, near the end-users and IoT sensors. For example, video could be processed in mini-datacenters deployed throughout the city, e.g., at bus stops. This load distribution reduces the risk of the ever-growing number of IoT devices overloading the data center. This thesis studies how to reorganise the process execution in this decentralised fashion, where processes must dynamically adapt to the volatile edge environment filled with moving devices. Namely, connectivity is intermittent, so decision-making and planning need to involve factors such as the movement trajectories of mobile devices. We examined this issue in simulations and with a prototype for Android smartphones. We also showcase the STEP-ONE toolset, allowing researchers to conveniently simulate and analyse these issues in different realistic scenarios, such as those in a smart city.
Kirjeldus
Märksõnad
Internet of things, sensor networks, information technology