Optimalno raspoređivanje dijelova videosignala kod prilagodljivoga stvarnovremenskoga strujanja u hibridnim mrežama ravnopravnih čvorova

doctoral thesis

• Title (english): Optimal chunk scheduling in adaptive live video streaming in hybrid peer to peer networks
• Author: Melika Mešković; MBZ: -333383
• Mentor: Mladen Kos (mentor)
• Committee member: Mladen Kos (predsjednik povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; (Department of Telecommunications); Zagreb
• Defense date and country: 2018, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing; Information Systems
• Universal decimal classification (UDC): 654 - Telecommunication and telecontrol

Abstract

Strujanje visokokvalitetnog videosignala putem hibridnih mreža ravnopravnih čvorova suočava se s brojnim izazovima uključujući: ograničen kapacitet otpreme čvorova, visoka heterogenost prijamnika u smislu propusnog opsega prijma, razlučivosti zaslona, kapaciteta procesora i visoki omjer odlazaka kao posljedica stalne promjene populacije čvorova. Rješavanje navedenih problema, u cilju nadoknade manjka resursa, ne zahtijeva samo povećanje kapaciteta čvorova s uvođenjem dodatnih poslužitelja, već i uvođenje novih metoda za kodiranje i distribuciju višemedijskog sadržaja i razvoj algoritama za optimalno korištenje raspoloživih sredstava. Iz tog je razloga pretpostavljena primjena skalabilnoga kodiranja videosignala kako bi se osigurala kvaliteta stvarnovremenoga strujanja videosignala u hibridnim mrežama ravnopravnih čvorova. Time se nastoje riješiti izazovi učinkovitom uporabom resursa čvorova, lakim prilagođavanjem višemedijskog sadržaja u cilju podrške prijamnika s različitim resursima i zahtjevima i brzom prilagodbom mreži i dinamici čvorova. Dat je prijedlog modela optimalnog raspoređivanja dijelova videosignala temeljenog na prioritizaciji tih dijelova, čiji je krajnji rok reprodukcije najbliži. Kako navedena prioritizacija nije dovoljna za postizanje apsolutne kvalitete usluge, uvedeno je i identificiranje međuovisnosti slojeva dijelova videosignala. Za određivanje prioriteta izračunava se i težinska vjerojatnost prijma dijela videosignala od susjednih čvorova. Predloženo raspoređivanje dijelova videosignala predstavljeno je u obliku NP-teškog problema pretvaranjem istog u problem naprtnjače, koji je problem kombinatoričke optimizacije. Za rješenje problema je primijenjen metaheuristički algoritam temeljen na tabu pretraživanju zbog sposobnosti brzog pronalaska optimalnog rješenja. Verifikacija predloženog postupka izvedena je simulacijskim eksperimentima u OMNeT++ simulatoru. Primjenom simulacijske metode izvršeno je vrednovanje učinkovitosti i usporedba predloženog pristupa s postojećim Random, Round Robin, Local Rarest First postupcima raspoređivanja dijelova videosignala na primjerima različitog broja dostupnih slojeva videosignala u stvarnom vremenu. Dokazano je kako će predloženo dinamičko raspoređivanje dijelova videosignala osigurati kritične parametre kvalitete usluge za stvarnovremeno strujanje videosignala, ispod standardom definiranih vrijednosti.

Abstract (english)

Provision of high quality video streaming over hybrid peer to peer networks is being faced with number of challenges that include: limited upload bandwidth of peers, high heterogeneity of peers in terms of bandwidth download capacity, display screen resolutions, processing power and high peer churn rate as a consequence of constantly changing peer population. Facing these challenges, so that the lack of resources can be compensated, does not only require an increase in capacity of the peers with introduction of additional servers but also requires introduction of additional methods for encoding and distribution of multimedia content together with the development of algorithms and protocols intended for optimal usage of available resources. For that purpose scalable video coding is applied so that the quality of live video streaming in hybrid peer to peer network can be assured. With this approach, challenges in providing live video streaming are tackled by efficient use of peer resources, effortless adaptation of multimedia content to network conditions in order to support heterogeneous user preferences and capabilities and peer dynamics. An approach has been proposed for optimal chunk scheduling in live video streaming based on chunk prioritization whose reproduction deadline is nearest. As stated prioritization is not sufficient enough to ensure the quality of service parameters, identification of multi-layer dependencies has also been introduced. During the process of prioritization, a weighted probability of receiving the chunk in due time is also taken into consideration. Proposed chunk scheduling of video stream is presented in a NP-hard problem turning the same into a knapsack problem which is a combinatory optimization problem. For the purpose of solving the stated problem a metaheuristic algorithm based on a taboo search has been introduced due to its capabilities of finding the optimal solution quickest. Verification of the proposed approach has been done using simulation experiments in the OMNET++ simulator. With the application of the simulation method an evaluation of the proposed approach efficiency was conducted together with a comparison with other approaches Random, Round Robin, Local Rarest First for chunk scheduling of live video stream based on the examples with varying number of available layers. It has been proven that with the application of the proposed approach for dynamic chunk scheduling in a live video stream the service will ensure that all quality of service parameters will be below maximum defined values by the service standard.

• Keywords: Hibridne mreže ravnopravnih čvorova; stvarnovremeno strujanje; dio videosignala; kvaliteta usluge; skalabilno kodiranje videosignala; raspoređivanje dijelova videosignala.
• Keywords (english): Hybrid peer to peer networks; live streaming; chunk; quality of service; scalable video coding; chunk scheduling algorithm.
• Language: croatian
• URN:NBN: urn:nbn:hr:168:216143
• Study programme: Title: Electrical Engineering and Computing; Study programme type: university; Study level: postgraduate; Academic / professional title: Doktor znanosti elektrotehnike i računarstva (Doktor znanosti elektrotehnike i računarstva)
• Type of resource: Text
• Extent: str. 155 ; 30 cm
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2019-04-01 14:21:08