diplomsko delo
Abstract
V diplomskem delu se ukvarjamo z reševanjem problema odpovedi v enojni točki v virtualiziranih strežniških sistemih. Ta problem rešujemo po teoretičnih priporočilih snovanja sistemov, odpornih na napake. Za implementacijo smo izbrali sistem VMware vSphere SMP-FT. V praktičnem delu diplomske naloge ta sistem vzpostavimo in implementiramo v testno okolje. Vzpostavljeno rešitev preizkusimo s simuliranjem okvare primarnega hipervizorja, pri čemer preverjamo, da ni prišlo do izpada virtualnega stroja. Prav tako natančno analiziramo vpliv mehanizma, ki zagotavlja odpornost na napake, na zmogljivost delovanja virtualnih strojev. Ugotovimo, da se pri uporabi mehanizma vSphere SMP-FT, ki uporablja priporočeno pasovno širino 10 Gbps za dnevnik, zmogljivost delovanja zmanjša 10–30 % glede na zmogljivost virtualnega stroja, ki takšnega mehanizma ne uporablja. Če uporabimo dnevnik s pasovno širino 1 Gbps, izmerimo drastično poslabšanje zmogljivosti delovanja (tudi do 90 %). Z množico eksperimentov potrdimo, da je mehanizem odpornosti na napake ustrezna rešitev za najbolj kritične dele strežniške infrastrukture. Pogoj je, da uspešno zadostimo vsem predpisanim zahtevam za implementacijo sistema, odpornega na napake, in da ni večjih omejitev pri dodeljevanju strojnih virov.
Keywords
virtualizacija;virtualni stroji;odpornost na napake;enojna točka odpovedi;VMware vSphere;diplomske naloge;
Data
Language: |
Slovenian |
Year of publishing: |
2019 |
Typology: |
2.11 - Undergraduate Thesis |
Organization: |
UM FERI - Faculty of Electrical Engineering and Computer Science |
Publisher: |
[R. Mori] |
UDC: |
004.946:004.416.2(043.2) |
COBISS: |
22903062
|
Views: |
479 |
Downloads: |
68 |
Average score: |
0 (0 votes) |
Metadata: |
|
Other data
Secondary language: |
English |
Secondary title: |
Fault tolerance in virtual machines |
Secondary abstract: |
This thesis focuses on solving the problem of a single point of failure in virtualized server infrastructure. We solve this problem following the theoretical recommendations of designing fault tolerant systems. VMware vSphere SMP-FT was selected for implementation. In the practical part of this thesis we established and implemented this system in a test environment. We test the established solution by simulating a fault in the primary hypervisor, verifying that the virtual machine hasn’t failed. We also carefully analyze the impact of fault tolerant mechanism on the performance of virtual machines. When using the vSphere SMP-FT mechanism at the recommended 10 Gbps logging bandwidth, we find the performance is reduced by 10-30%, by comparison with a virtual machine without fault tolerance. When using a 1 Gbps logging bandwidth, we measure a drastic performance drop (up to 90%). With a multitude of experiments, we confirm that fault tolerance mechanism is an appropriate solution for the most critical parts of server infrastructure. On condition that we successfully meet all the prescribed requirements for implementing a fault tolerant system and that there are no major restrictions on hardware resource allocation. |
Secondary keywords: |
virtualization;virtual machines;fault tolerance;single point of failure;VMware vSphere; |
Type (COBISS): |
Bachelor thesis/paper |
Thesis comment: |
Univ. v Mariboru, Fak. za elektrotehniko, računalništvo in informatiko, Računalništvo in informacijske tehnologije |
Pages: |
IX, 37 str. |
ID: |
11213454 |