This dissertation thesis deals with the architecture, benchmarking, optimization and
implementation of virtual distributed computing systems. Large distributed systems rep-
resenting the performance background of all modern cloud computing architectures have
become a very hot topic at present. This dissertation thesis o ers an introduction into
modern technologies showing a rapid impact on the system performance. One of them is
virtualization technology, whose real e ciency was a standalone part of the research. Large
distributed systems consume a huge amount of electric power, therefore their optimization
was also discussed. New ideas originated from the research were incorporated into the pro-
posal of a new distributed system called Centr aln mozek univerzity (CMU). This system
is able to manage and automatically optimize large virtualization infrastructures. It is also
accessible to the end-users. This system is currently used in teaching many subjects and
became a prototype of an educational system of new generation.
In particular, the main contributions of the dissertation thesis are as follows:
1. Methodology, design and implementation of a new software benchmark utility able
to measure virtualization e ciency of the distributed architecture.
2. New approach in a possible migration of the entire distributed systems between
various data centres.
3. New distributed algorithm for virtual machines consolidation, rapidly reducing the
energy consumption.
4. Design, description and implementation of the distributed virtualization system CMU.
en
dc.description.abstract
Diserta cn pr ace se zaob r a architekturami, testov an m, optimalizac a implementac
virtu aln ch distribuovan ych syst em u. Rozs ahl e distribuovan e syst emy, kter e stoj v pozad
v sech modern ch cloudov ych architektur, jsou v sou casn e dob e skute cn e zhav ym t ematem.
Pr ace obsahuje teoretick y uvod do modern ch virtualiza cn ch technologi , kter e rapidn e
ovliv nuj v ypo cetn v ykon. Efektivita virtualiza cn ch technologi byla jedn m z kl cov ych
t emat vlastn ho v yzkumu. Rozs ahl e v ypo cetn infrastruktury spot rebov avaj obrovsk e
mno zstv elektrick e energie, p ri cem z optimalizace spot reby byla dal s m p redm etem v yzkumu.
Inovativn my sleny, kter e vznikly ve f azi v yzkumu, byly integrov any do n avrhu a im-
plementace nov eho distribuovan eho virtualiza cn ho syst emu, kter y nese n azev Centr aln
mozek univerzity (CMU). Tento syst em je schopen pln e automatizovan e spravovat a op-
timalizovat distribuovan e infrastruktury pro virtualizaci. Syst em dok a ze zprost redkovat
p r stup pro vyu zit sv ych v ypo cetn ch prost redk u koncov ym u zivatel um. Takov ato kon-
cepce syst emu je sou casn e prototypem v yukov eho syst emu nov e generace, ve kter em je
vyu zita architektura datov ych center pro v yuku univerzitn ch p redm et u.
cze
dc.language.iso
en
en
dc.subject
virtualization
en
dc.subject
distributed computing systems
en
dc.subject
cloud computing
en
dc.subject
live migration
en
dc.subject
green computing
en
dc.title
Virtual Distributed Computing Systems and Their Applications
en
dc.type
disertační práce
cze
dc.description.department
Katedra počítačů
theses.degree.discipline
Informatika a výpočetní technika
theses.degree.grantor
České vysoké učení technické v Praze. Fakulta elektrotechnická. Katedra počítačů