Predviđanje uporabe resursa u sustavima temeljenim na programskim komponentama

doctoral thesis

• Title (english): Resource usage prediction in component-based software systems
• Author: Marin Orlić
• Mentor: Mario Žagar (mentor)
• Committee member: Danko Basch (predsjednik povjerenstva)
• Committee member: Mario Žagar (član povjerenstva)
• Committee member: Ivica Crnković (član povjerenstva)
• Committee member: Igor Čavrak (član povjerenstva)
• Committee member: Mario Kušek (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; (Department of Control and Computer Engineering); Zagreb
• Defense date and country: 2010-11-04, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing
• Universal decimal classification (UDC): 004 - Computer science and technology. Computing. Data processing

Abstract

Ugradbeni sustavi kao posebna skupina računalnih sustava imaju sve širu primjenu. Složenost razvoja ugradbenih sustava postavlja mnoga pitanja - učinkovitost uporabe resursa samo je jedno od njih, no ključno za razvoj kvalitetne programske potpore. Razvoj u smjeru integracije funkcionalnosti više uređaja u jedan i potreba za bržim i jeftinijim razvojem samo će povećati potrebu za metodama i postupcima razvoja učinkovitih sustava. U razvoju sklopovlja uređaja uobičajena je analiza ponašanja sustava u ranoj fazi razvoja, što nije slučaj u razvoju programske potpore. Disertacija daje pregled komponentnog modela i modela ponašanja, predlaže model izvedbenog okruženja programske potpore i mogućnosti integracije različitih modela programskih komponenata. Navedeni modeli nužni su kao pretpostavka za razvoj postupaka predviđanja promjene i uporabe resursa sustava temeljenih na programskim komponentama. Kao ogledni komponentni model uzet je komponentni model za ugradbene sustave ProCom razvijen na Sveučilištu Mälardalen, Švedska i pripadni jezik (model) za opis ponašanja REMES koji je još u razvoju. Postupci predviđanja promjene i uporabe resursa zasnovani su na simulacijskom modelu dobivenom integracijom različitih modela u cjeloviti model sustava. Ogledni alati dokazuju mogućnosti primjene postupaka za analizu ponašanja sustava u ranoj fazi razvoja, kada su dostupni samo modeli sustava bez podataka o konačnoj izvedbi sustava.

Abstract (english)

Embedded systems, as a special class of computer systems have ever wider applications. Complexity of software development for embedded system raises many questions. Resource usage efficiency is just one of them, but nonetheless an essential one in the process of development of quality software. Development directions show future embedded systems will integrate functionality of many devices into one. These also indicate an increasing need for new methodologies and development processes resulting in development of efficient systems. Hardware development process is an example of system behavior analysis applied early on in development stages, a feat often missing in the process of software development. The question this thesis attempts to answer is - is it possible to predict future system's resource usage while still in early design phase? Thesis gives an overview of component and behavior models, proposes a profile of an executive environment and discusses the possibilities of integration of various component models. Models are a necessity for the process of resource usage analysis of component-based software. ProCom component model and REMES behavior model developed at the Mälardalen University, Sweden, were chosen as reference models. Resource usage models used in this thesis allow specification of constraints and resource usage in real-time systems. Resource usage prediction methods developed are based on simulation models resulting from the integration of various component models into integrated system model. Simulation was chosen as a method applicable in the early design phase when only partial design information is available, which hinders classic analysis, e.g. formal methods. Proposed simulation methods continues similar work with priced timed automata and hybrid system simulators, and is designed to adapt to REMES behavior model structure. As a part of a larger integrated development environment effort around ProCom component model, tooling was developed to demonstrate the possibilities of applying these methods in analysis of system resource-wise behavior in early design phase, with only limited system models and no available deployment models. The contribution of this thesis is in the following: execution environment model for component-based systems, describing system resources and resource limitations; an integral system model, generated from partial structure and behavior models and execution environment models; methods to generate the integral system model from available models; behavior simulation methods that can be used to predict resource usage; tooling and a design environment that implement and automate model generation and analysis. Proposed methodologies have been compared with state-of-the-art formal analysis tools and demonstrate comparable results. Future work should be directed towards improvement of end-user tools to enable further testing of proposed methods. Both underlying component and behavior models chosen as a basis for this work are still in development, and the simulation and model merge methods will likely need to adapt to changes with respect to changes in underlying models.

• Keywords: programsko inženjerstvo temeljeno na programskim komponentama; modeliranje ponašanja; analiza utroška resursa; opis ponašanja u vremenskoj domeni; komponentni modeli; modeli ponašanja; razvoj vođen modelom; ugradbeni sustavi
• Keywords (english): component-based software engineering; behavior modeling; resource usage analysis; timing behavior; component models; behavior models; model-driven architecture; embedded systems
• Language: croatian
• URN:NBN: urn:nbn:hr:168:154883
• Project: Number: 036-0361959-1965; Title: Programsko inženjerstvo u sveprisutnom računarstvu; Leader: Mario Žagar; Jurisdiction: Croatia; Funder: MZOS; Funding stream: ZP
• Study programme: Title: Computer Science; Study programme type: university; Study level: postgraduate; Academic / professional title: Doktor znanosti (Doktor znanosti)
• Catalog URL: http://lib.fer.hr/cgi-bin/koha/opac-detail.pl?biblionumber=35434
• Type of resource: Text
• Extent: 142 str ; 30 cm
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2019-06-12 12:34:17