COEN 241 HW 1 System vs OS Virtualization

COEN 241 HW 1
System vs OS Virtualization
Learning Objectives
The aims of the assignment include the following learning objectives:
● Understanding the real-world use case of two virtualization techniques / technologies:
○ System Virtualization via QEMU
○ OS Virtualization via Docker
● Understand the operations related to each technologies
● Understand and demonstrate the performance differences of each technology using
benchmarks and finally compare the results in a comprehensive report.
Environment Setup
You are free to use your personal computer for this assignment. The set of requirements for
using your personal computers is as follows:
● x64 CPU with at least 2 cores or a Mac Apple Silicon
● 4 GB memory
● 20 GB free disk space
● OS: Windows, Linux (Ubuntu preferred) or Mac OS X
If you are using a personal computer that does not support such requirements, please do reach
out to me, so that we can discuss some options.
System Virtualization (QEMU) Setup
What is QEMU?
QEMU is a free and open-source hypervisor. It emulates the machine's processor through
dynamic binary translation and provides a set of different hardware and device models for the
machine, enabling it to run a variety of guest operating systems. It is more lightweight than
virtualbox and allows near-native speed (when using QEMU with KVM on Linux). QEMU
supports various types of CPUs, such as x64 and ARM, and also supports integrations with
various host operating systems. It is less user-friendly and requires more effort to setup.
Installing QEMU
For All Operating Systems
For this assignment, you will be installing a Ubuntu guest virtual machine. Thus you must
download the appropriate Ubuntu 20.04 or 16.04 server ISO image from one of the links below:
● x64 CPU on Linux or Mac: Ubuntu 20.04 Server
● x64 CPU on Windows: Ubuntu 16.04 Server
● ARM (Apple Silicon): Ubuntu 20.04 Server for ARM
From here, the installation instruction varies between which OS you are running.
Optional helpful tips:
● You can set -nographic to run directly in the terminal.
● Another option is to SSH from your host machine to the VM by adding the following flags
to enable port mapping from 8888 to 22 and run the SSH command shown below.
● -netdev user,id=net0,hostfwd=tcp::8888-:22
● $ ssh -p 8888 localhost
● Do the above only if you have successfully launched a VM!
Linux Installation instructions
QEMU is the easiest to install if you have Ubuntu. You can install qemu (under Ubuntu) simply
by typing the following command in the terminal:
$ sudo apt-get install qemu
You can then create the QEMU Image by running the following command.
$ sudo qemu-img create ubuntu.img 10G -f qcow2
Given the image, install the VM using the command below (which takes the iso file as a “cdrom”
and the qemu image as a “hard disk”):
$ sudo qemu-system-x86_64 -hda ubuntu.img -boot d -cdrom
./[UBUNTU_SERVER_ISO_FILE_NAME] -m 2046 -boot strict=on
For each of these commands, please refer to the QEMU manual for more details [1].
Mac OS X (x64) Installation instructions
The instructions to install QEMU on the x64 Macs can be found in the following link
https://wiki.qemu.org/Hosts/Mac. I suggest using the homebrew method as it is the easiest. You
can find homebrew installation instructions in [2], if it is not already installed in your machine.
Once QEMU is installed, you can follow a similar procedure as the Linux installation to create a
QEMU image and install the VM. Things to watch out for:
● Make sure to update homebrew
● Try 20.04, if it doesn’t work 16.04 may work.
Mac OS X (Apple Silicon) Installation instructions
Given that Apple Silicon is a relatively new processor, there are not many resources available to
simply install QEMU with a single command line. Once you install qemu using brew as similar to
above, first make sure you have version 6.2.0+, because that is the minimum version that
supports Apple Silicon. If not, you should upgrade to 6.2.0.
Also, make sure you are not running your terminal via Rosetta. To check, type in arch in the
terminal to make sure it is arm64, not i386.
Here is a link that may help, but given that the guide is for QEMU 5.2, it won’t work as is with
version 6.2.
https://gist.github.com/citruz/9896cd6fb63288ac95f81716756cb9aa#gistcomment-3629192
Here is a set of commands that may work:
● $ qemu-system-aarch64 \
-accel hvf -cpu cortex-a57 -M virt,highmem=off -m 2048
-smp 2 \
-drive
file=/usr/local/share/qemu/edk2-aarch64-code.fd,if=pflash,format
=raw,readonly=on \
-drive if=none,file=ubuntu.img,format=qcow2,id=hd0 \
-device virtio-blk-device,drive=hd0,serial="dummyserial" \
-device virtio-net-device,netdev=net0 \
-netdev user,id=net0 \
-vga none -device ramfb \
-cdrom ubuntu-20.04.4-live-server-arm64.iso \
-device usb-ehci -device usb-kbd -device usb-mouse -usb \
-nographic
Note that the VM image file names may be different based on which version you have
downloaded. Make sure you download the ARM64 image.
Windows Installation Instructions
Running QEMU on Windows is NOT RECOMMENDED, as most servers are Linux/Unix based
and most open source virtualization works best with Unix based OSes. However, it is certainly
possible. Here is an article that describes the entire process in detail in [3], which requires you
to be able to change some Window environments variables.
● You may need to install VcXsrv to connect to the VM instance. Once you have
connected Ubuntu to the display, follow the Linux instructions on the homework.
Last Resort!
For All OSes, if all things above do not work, you can use VirtualBox (for Windows & Linux) and
UTM (https://mac.getutm.app/). Here are some additional requirements if you choose to take
this path.
● List the steps you have tried and screenshots of the errors if possible.
● Run VirtualBox and UTM using the command line argos
○ VirtualBox:
https://www.oracle.com/technical-resources/articles/it-infrastructure/admin-manag
e-vbox-cli.html
○ UTM does not have a CLI currently. However, you can export the QEMU
command it runs: open VM Configuration, click the QEMU tab and then Export
QEMU arguments…
● Document your command line configurations for Virtualbox or UTM
OS Virtualization (Docker) Setup
What is Docker?
Docker is a set of platforms that enable OS-level virtualization to deliver software in packages
called containers. Docker is the most popular container management platform.
Installing Docker
While we will be installing Docker Desktop (for Windows and Mac), we will be interacting mostly
with Docker CLI. Docker Desktop installation will include Docker Engine, Docker CLI client,
Docker Compose and more. You can learn more about the Docker CLI in [5]. I have included the
corresponding links to find the installation instructions for each type of OSes. Installing Docker
should be much simpler than installing QEMU.
Linux Installation instructions
Note: Ubuntu does not come with Docker Desktop, so we would just need to install the Docker
Engine and the Docker CLI separately.
● https://docs.docker.com/engine/install/ubuntu/
Mac OS X Installation instructions
● x64 CPU: https://docs.docker.com/desktop/mac/install/
● ARM (Apple Silicon):: https://docs.docker.com/desktop/mac/apple-silicon/
Windows Installation instructions
● https://docs.docker.com/desktop/windows/install/
Installing sysbench
Sysbench[4] is an open-source and multi-purpose benchmark utility that evaluates the
parameter features tests for CPU, memory, I/O and much more. We will use sysbench to
understand the performance characteristics of each of the virtualization technologies.
https://github.com/akopytov/sysbench/issues/140
For Ubuntu Guest VM
Sysbench installation on Ubuntu is very simple. Just run the following commands:
$ sudo apt update
$ sudo apt install sysbench
Note: Do not be confused with the password for the guest VM.
For Docker
For Docker, you will be creating your own image by adding sysbench to a base image. Based
on your choice of Ubuntu version and your architecture, you will choose a base image in Docker
Hub. You can find the correct Docker docker images in https://hub.docker.com/_/ubuntu. Make
sure to match the ubuntu version and your chip! Very important!
Once the base image is downloaded, you must install sysbench similar to how you installed it on
the VM. Note that the base image is quite light, so it may not contain simple commands like
sudo or ping.
You must describe the commands you ran to create your image, your final image ID, and
the screenshot of your image history!
Note on Sysbench Versions!
The sysbench version on your linux machine and the Docker container may differ, as sysbench
version is updated continuously. Therefore, the output values may seem like they are different.
To eliminate this issue, make sure you have the same version of sysbench on both experiments.
What to Experiment?
Given the sysbench tool, we can have some basic ideas about the system performance by
running it against the system under test. In assignment, we will focus on the cpu and fileio test
modes (Section 4.1 and 4.5 in [4]) of Sysbench. The following instructions present some ideas
about how to conduct a meaningful measurement. Please read them carefully, and then produce
the measurement results. If the results are incorrectly generated and/or can not be well justified,
you will not be able to get a full credit for this assignment.
● Test different arguments of qemu
○ Some examples: -m, -smp, -accel and more
○ Be sure to document your experiments and why you used them.
● You need to figure out all the test cases (under the cpu and fileio test modes).
● For each test case, you need to figure out the“right” parameters. Some examples are,
1. Choose the parameter for “–cpu-max-prime” to ensure your test case won’t end
in a short time period (too small) or will never end (too large).
a. Note that if your experiment always runs for 10 seconds, change the time
parameter and report events/sec.
2. Choose the –time parameter accordingly
3. Decide the file size (e.g., –file-total-size) for Sysbench fileio test mode.
● Usually, it’s reasonable to make a test case lasting for at least 30 seconds, but for such
test cases, please justify your configurations for doing so.
● You need to repeat the sysbench measurement at least 5 times for each test case, and
report the average, min, max and std. values of your results (the Sysbench will report
some user-level performance data, e.g., total time).
● To reduce test variation, you need to test each case under a similar test environment.
○ For the fileio test, after one test, the operating system will cache the accessed
files. If we don’t manually drop such cache, the following tests will finish much
faster, as most I/O data will be served directly from the 1 memory cache instead
of disks. You have to manually drop such cache in the Host using the commands
(in superuser or administrator mode):
■ For Linux:
echo3 > /proc/sys/vm/dropcaches
■ For Mac OS:
sync && sudo purge
■ For Windows (Windows, this is not readily available, thus you must
download the tool from here):
sync
● For each of the experimental runs, create a bash script to automate the experiment.
● Please well organize your experimental setup, configurations, and data (using figures
and/or tables); and report them.
● (Optional) If using Virtualbox or UTM, check what you should add in the ‘Last Resort!’
section.
Submission Instruction
Create a submission for the assignment on the Camino containing the following two pieces of
information:
1. a URL such that the command ‘git clone URL’ would download your repository
(assuming that the repository has been shared with the requesting user);
2. a git commit ID within your repository that you want markers to assess that contains the
following.
a. The project report in PDF format
b. Shell scripts to run your experiments
3. You will reuse this repository for the entire class, so do not name it something like
COEN241-HW1.
Repository Requirements
● You are expected to work on your assignment within a git repository. The easiest way to
acquire URLs that reach your repository is to push your work to a (free) cloud-based git
service, such as GitHub, Bitbucket, GitLab, or Altitude (a GitLab server run locally within
the Department of Computer Science that you can access). The aforementioned cloud
services allow you to create private repositories.
● You will create a SINGLE repository for all homework assignments and projects,
so make sure you create a folder called ‘hw1’ to store all your HW 1 files.
● You must invite teaching staff to collaborate on your repository, so that they can access
the URL that you submit. Please send an invitation to sean.choi@scu.edu
● Do not apply git history-rewriting operations to your repository: it is important that the
markers can see how your commits completed your assignment, and how you worked
on the assignment over time.
Rubric / What to Write on the Report
● Your report should be a single PDF file included in the github repository. Most of the
grade on this assignment is based on this report.
● You should not only follow the instructions to do your assignment, but also fully
understand what you are doing
● The report should include (but not limited to):
○ Detailed configurations (CPU, Mem, etc...) of your experimental setup: 5 points
○ Present main steps to enable a QEMU VM. In addition, please present the
detailed QEMU commands, and VM configurations: 10 points
○ Present main steps to enable the Docker container. This must show your steps
in creating your own image and your image history! Do not copy any
classmates image. In addition, please describe the operations you use to
manage Docker containers (and some other operations which you think are also
important): 10 points
○ Proof of experiment. Include screen snapshots of your Docker and QEMU
running environments for each experiment: 5 points
○ Present how you conduct your measurements in three different scenarios for
each virtualization technology: 20 points
○ Shell scripts for running the experiment: 20 points
○ Present how you use performance tools to collect performance data. For CPU
utilization, you should at least divide them into two parts including user-level and
kernel-level. For I/O, you should present I/O throughput, latency, and disk
utilization: 10 points
○ Presentation and analysis of the performance data: 10 points
○ Understandability/Neatness of your report: 10 points
○ Git Repository Information: 5 points
○ Extra Credit: Provide the correct Vagrant and Docker files for your VM and
Docker Container. 20 points
About the Extra Credit
For cloud deployment, automation is critical. Thus there are tools that enable developers to
easily store the configurations of their VMs & Containers and launch them on the cloud. One
such tool is Vagrant and the other is Docker File. Writing the configurations for these platforms
requires practice and may take multiple tries before getting them right. However, if you are
ambitious, you can try to write both Vagrant and Docker File for this assignment for future uses
(and be rewarded hefty for it!). Note that Vagrant was created to be used with VirtualBox, so it
may be tricky to run and debug the Vagrant file with QEMU. My suggestion is to do the
assignment using QEMU, but build the Vagrant file using Virtualbox. If you are ambitious, you
can try running Vagrant with QEMU as well! However, note that the extra credit is the same
regardless of which hypervisor you decide to test the Vagrant file on.
Here are some references to help you write Vagrant and Docker Files
● Writing Vagrant File https://www.vagrantup.com/docs/vagrantfile
● Writing Docker File
https://docs.docker.com/develop/develop-images/dockerfile_best-practices/
For those of you who want to run Vagrant with QEMU, this would be the first step to try.
● https://computingforgeeks.com/using-vagrant-with-libvirt-on-linux/
● https://github.com/ppggff/vagrant-qemu
References
[1] QEMU Documentation: https://qemu.weilnetz.de/doc/6.0/
[2] Homebrew: https://brew.sh/
[3] How do I run QEMU on Windows: https://linuxhint.com/qemu-windows/
[4] Sysbench manual: https://imysql.com/wp-content/uploads/2014/10/sysbench-manual.pdf
[5] Docker CLI Reference: https://docs.docker.com/engine/reference/commandline/cli/
[6] Docker exec: https://docs.docker.com/engine/reference/commandline/exec/
[7] Docker Tutorials: https://docker-curriculum.com/