The capacity of monitoring events occuring inside virtual machines is important for diagnostic purposes. The OS running withing the VM, however, ignores to be running in a virtualized environment. Among the many issues in exporting VM events outside the virtualized environment, the synchronization between events inside the VM and outside the VM is of paramount importance. The purpose of this work is to develop a tool that permits to infer the accuracy of the synchronization when tracing both host and guest(s). For this purpose, an approach based on the analysis of the events is presented and specific sequences of them are investigated to determine if they can be used to establish a synchronization point. Also, state-of-the-art synchronization algorithms are analyzed and a comparison among them is presented. To better understand this particular type of tracing, it is necessary to introduce general concepts about the Linux kernel, the virtualization and how to make tracing in this enviroment. Then, how a merged trace looks, how it is obtained and how we can infer useful information from it. Finally, the "event-based'' approach and how to properly automate it by using KernelShark 2 libraries.
Metodologia per la Valutazione dell'Accuratezza nella sincronizzazione tra Hypervisor e tracce di Macchine Virtuali
DE VENUTO, STEFANO
2020/2021
Abstract
The capacity of monitoring events occuring inside virtual machines is important for diagnostic purposes. The OS running withing the VM, however, ignores to be running in a virtualized environment. Among the many issues in exporting VM events outside the virtualized environment, the synchronization between events inside the VM and outside the VM is of paramount importance. The purpose of this work is to develop a tool that permits to infer the accuracy of the synchronization when tracing both host and guest(s). For this purpose, an approach based on the analysis of the events is presented and specific sequences of them are investigated to determine if they can be used to establish a synchronization point. Also, state-of-the-art synchronization algorithms are analyzed and a comparison among them is presented. To better understand this particular type of tracing, it is necessary to introduce general concepts about the Linux kernel, the virtualization and how to make tracing in this enviroment. Then, how a merged trace looks, how it is obtained and how we can infer useful information from it. Finally, the "event-based'' approach and how to properly automate it by using KernelShark 2 libraries.File | Dimensione | Formato | |
---|---|---|---|
901800_devenuto901800.pdf
non disponibili
Tipologia:
Altro materiale allegato
Dimensione
2.45 MB
Formato
Adobe PDF
|
2.45 MB | Adobe PDF |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.14240/130914