Comparative Study of the Support of Academic Clouds Based on Apache

CloudStack and Proxmox VE Platforms

Vasyl P. Oleksiuk

1,2 a

, Olesia R. Oleksiuk

3 b

and Oleg M. Spirin

2,4 c

Ternopil Volodymyr Hnatiuk National Pedagogical University, 2 M. Kryvonosa Str., Ternopil, Ukraine

Institute for Digitalisation of Education of the National Academy of Educational Sciences of Ukraine,

9 M. Berlynskoho Str., Kyiv, 04060, Ukraine

Ternopil Regional Municipal Institute of Postgraduate Education, 1 V. Hromnytskogo Str., Ternopil, 46027, Ukraine

University of Educational Management, 52A Sichovykh Striltsiv Str., Kyiv, 04053, Ukraine

ﬁ

Keywords:

Cloud Computing, Academic Cloud, Apache CloudStack, Proxmox VE, Google Workspace.

Abstract:

Cloud technologies provide users with efﬁcient and secure tools for data management, computing, storage and

other services. The article analyzes the projects for the introduction of cloud technologies in education and

identiﬁes the main advantages and risks in creating a cloud infrastructure for the university. Such startups

contribute to the formation of a new paradigm of education. It involves the virtualization of education, the

introduction of mobile and blended learning, ie the combination of cloud computing with modern learning

concepts. In this paper, we compare some aspects of deployment and supportin our experience in improving

the academic cloud for the training of a bachelor’s degree in computer science. This is through the integration

of the Proxmox VE platform into existing computing power by deploying the Proxmox VE system. In the

study, we reveal some technical and methodological aspects of the organization of the educational process

using this corporate cloud platform. The scheme of the organization of physical components of cloud infras-

tructure (nodes, virtual networks, routers, domain controller, VPN-server, backup system of students’ virtual

machines) is given. All characteristics of this environment and possibilities of their application are studied.

1 INTRODUCTION

Today, many universities are creating their own cloud-

based learning environments (CBLE). Although there

is currently no single concept for CBLE, scientists

understand it as similar concepts (Glazunova and

Shyshkina, 2018; Korotun et al., 2020; Bykov et al.,

2020; Vakaliuk, 2021; Spirin et al., 2022). In general,

it can be understood as an IT system consisting of

cloud services and providing learning mobility, group

collaboration of teachers and students to achieve edu-

cational goals (Lytvynova, 2018).

As the analysis of the literature shows, many uni-

versity CBLE are usually deployed according to the

hybrid model (Shyshkina, 2016; Markova et al., 2019;

Vorozhbyt, 2017; Vakaliuk et al., 2021). One of the

most important components in the structure of this en-

vironment is the private academic cloud (Glazunova,

https://orcid.org/0000-0003-0789-0272

https://orcid.org/0000-0001-5659-4774

https://orcid.org/0000-0002-9594-6602

2015). It is now deployed according to the most

productive IaaS service model. A hybrid cloud is

a cost-effective way to solve the problem of insuf-

ﬁcient computing resources. The private academic

cloud allows the university to meet the peak demand

of students and faculty through a combination of local

infrastructure and one or more public clouds (Wang

et al., 2020).

Various commercial and free platforms are used

in universities to build private academic clouds (AL-

Mukhtar and Mardan, 2014; Ilin et al., 2016; Amiel

et al., 2020). A productive method of deploying pri-

vate academic clouds is to use solutions from lead-

ing cloud vendors such as Google, Microsoft, Ama-

zon and others. Google offers researchers, universi-

ties, faculties, faculty and students grants and loans

for teaching and research. In particular, leading Eu-

ropean educational institutions can access Google

Cloud within the Internet2 project (Internet2, 2023).

Unfortunately, these opportunities are not currently

available for our country (Ukraine). For example, Mi-

crosoft Educator Grant is a program designed specif-

Oleksiuk, V., Oleksiuk, O. and Spirin, O.

Comparative Study of the Support of Academic Clouds Based on Apache CloudStack and Proxmox VE Platforms.

DOI: 10.5220/0012064300003431

In Proceedings of the 2nd Myroslav I. Zhaldak Symposium on Advances in Educational Technology (AET 2021), pages 349-361

ISBN: 978-989-758-662-0

 2023 by SCITEPRESS – Science and Technology Publications, Lda. Under CC license (CC BY-NC-ND 4.0)

349

ically to provide access to Microsoft Azure to college

and university professors teaching advanced courses.

As part of this program, faculty teaching Azure in

their curricula are awarded subscriptions to support

their course (Microsoft, 2023). In general, these pro-

grams are very useful and productive. However, they

are usually provided on a temporary basis and there-

fore cannot completely replace the cloud-based IT in-

frastructure of universities and colleges.

Among the free platforms for cloud infrastructure

deployment, CloudStack, Openstack, Proxmox VE,

and Eucalyptus are the most suitable. Each of them

has its advantages and disadvantages. There are many

attempts to compare these platforms. The authors of

articles comparing such platforms state (Microsoft,

2023; Fylaktopoulos et al., 2016):

• OpenStack has large community, offers wide inte-

gration with storage, network and compute tech-

nologies, but is too complex to deploy and conﬁg-

ure;

• Eucalyptus the longest-standing open source

project is banking on its very tight technical ties

to Amazon Web Services (AWS). The platform is

conﬁgurable, but not very customizable;

• Proxmox is open source platform. It can provide

easy way to deployment cloud infrastructure. But

it is not very suitable as a platform for a private

cloud in the CLBE;

• CloudStack has well rounted GUI, can provide an

advanced cloud infrastructure, but it is very GUI

centric built on single Java core.

We have been deploying a private academic cloud

based on the Apache CloudStack platform for the

past 8 years. For the last 2 years we have been de-

ploying an academic cloud based on Proxmox VE as

well. Apache CloudStack deployment is housed in

the infrastructure building. It contains a management

server, 4 hosts, 4 primaries and 1 secondary storage.

We decided to use hypervisors instead of containers.

This is because the former are more versatile. In ad-

dition, the use of hypervisors is safer than containers.

To save computing resources, we have installed pri-

mary storage on the hosts. We used VLANs to dis-

tribute trafﬁc across individual networks. These net-

works can be allocated to groups or individual stu-

dents.

We have so far rented one refurbished physical

server based on a used Intel Xeon processor with 32

Gb of RAM for deployment of Proxmox VE. This was

also done because Ukraine is currently at war and it

is difﬁcult to say whether servers will be available as

they are in university buildings.

In general, our private academic cloud provides

(Fylaktopoulos et al., 2016):

• Development and Execution of student virtual

machines;

• Aggregation of computing resources of several

hosts;

• VM migration between repositories;

• VM connections to each other through guest net-

works;

• Launching VMs within other VMs;

• Integration with Active Directory;

• Distribution of student accounts according to their

academic groups.

There are many advantages and disadvantages to

using Apache CloudStack and Proxmox VE as an aca-

demic cloud platforms. There are some problems

happens in the process private academic cloud using.

The purpose of the article is to compare some

tasks in maintenance of academic clouds based on

Apache CloudStack and Proxmox VE platforms.

The following tasks are required to achieve the

goal of the research:

1. Analysis of maintenance tasks of academic clouds

in foreign and Ukrainian universities.

2. Deﬁnition of maintenance tasks of academic

clouds based on Apache CloudStack and Prox-

mox VE platforms.

3. Describe and compare methods for solving these

tasks.

We used a set of research methods such as theo-

retical (analysis of scientiﬁc, technical literature, ex-

perience; generalization of experience of using cloud

computing in education) and empirical (modelling,

designing, developing of scripts).

2 THE PRIVATE CLOUDS

MAINTENANCE TASKS

As the experience of cloud infrastructure maintenance

shows, this is an ongoing process. It requires constant

attention from engineers, network system administra-

tors, teachers, and student involvement. Scientists de-

scribe the experience of deploying an academic pri-

vate cloud, including determining the performance of

hypervisors and storage (Spirin et al., 2018). The

biggest challenge for researchers was the transition

from a prototype of an academic cloud to a produc-

tive one. In this context, they addressed the prob-

lem of load balancing, elastic hypervisors, security

AET 2021 - Myroslav I. Zhaldak Symposium on Advances in Educational Technology

350

threats. Storage backup tasks are also important for

such clouds.

Rao and Nayak (Rao and Nayak, 2014) offer the

concept of backup to cloud storage. They say that

both the cloud provider and the cloud consumers have

to take comprehensive steps to ensure appropriate

conﬁgurations, hardening of the CBLE, appropriate

design and development, appropriate interoperability,

and adequate testing.

Rusyn et al. (Rusyn et al., 2019) have developed

an effective method of deduplication and distribution

of data in cloud storage during the creation of back-

ups. Researchers have developed an intelligent sys-

tem for such deduplication and tested it.

Tian et al. (Tian et al., 2020) also studied cloud

data backup. The authors propose a scheme for data

separation and backup and encryption. They state that

their own scheme resolves the conﬂict between data

security and the survivability of the IT infrastructure

with the help of encrypted backup.

The Apache CloudStack cloud infrastructure re-

dundancy model developed by Paul Angus is very

useful for our study. It creates a vendor agnostic

API and UI in CloudStack for end users. The au-

thor’s Framework abstracts the speciﬁcs of solutions,

such that through the use of a plugin, a 3rd party

solution can deliver backup and recovery solutions

(Khmelevsky and Voytenko, 2015) .

Here are some tasks for servicing our sample pri-

vate academic cloud:

1. Making changes to the cloud infrastructure.

2. Creating and veriﬁcation of the academic cloud

performance model.

3. Designing and realization an academic clouds’

backup model.

Let’s consider these tasks in more detail.

3 MAKING CHANGES TO THE

CLOUD INFRASTRUCTURE

Maintenance of our developed cloud infrastructure in-

volves the implementation of tasks such as:

• changing the parameters of the components of the

cloud infrastructure – zones, clusters, hosts, stor-

ages;

• creating and routing of virtual networks for indi-

vidual groups or students;

• creating and modifying templates of compute of-

fering services that determine the performance of

VM;

• creating and modifying network offering service

templates such as VPN, DHCP, DNS servers,

Firewall, Load Balancer and others;

• creating projects for VM sharing by students.

This task also involves creating student and fac-

ulty accounts. We authenticate users of the academic

cloud from a centralized database – LDAP-directory

(Microsoft Active Directory). This approach makes

it possible to use single registration data to access

all hybrid IT infrastructure services. We used Cloud-

Stack domains to distribute students according to aca-

demic groups. Adding users to them is possible in

automatic (using links at the ﬁrst successful authenti-

cation) and manual mode. Unfortunately, due to the

incompatibility of our users’ logins with the Apache

CloudStack platform, we had to choose the manual

mode. To reduce the technical work involved in ﬁnd-

ing LDAP directory entries, we have created several

queries to ﬁlter user account data.

Proxmox VE also provides authentication based

on the LDAP directory (Active Directory). To do this,

cloud administrator need to specify some parameters

such as socket to connect to the domain controller,

user data for reading data from the directory, some

user attributes, ﬁlters to restrict access.

To log in to the system, the administrator can

add it manually, for example using the web interface.

However, it is possible to automatically synchronize

the directory with the pveum realm sync <realm>

command. Like Apache CloudStack domains, Prox-

mox VE can create groups. For these groups, cloud

administrator can deﬁne privileges for access to cloud

infrastructure.

Apache CloudStack users automatically have ac-

cess to their own VMs. This is good for students to

complete individual tasks. The creation and providing

permission to VMs isn’t a trivial task in the Proxmox

VE Platform. We offer two approaches to do this.

The ﬁrst involves the administrator creating copies of

VMs and giving students access to these VMs. Per-

missions may vary depending on the operations to be

performed with these VMs. For example, cloud ad-

ministrators can provide students with administrative

access to such VMs. Another approach involves stu-

dents cloning their own machines from the templates.

It requires providing some rights. They are displayed

as a table 1.

Creating permissions that match the last row

of the table requires deﬁning <students login> -

<VM ID> pairs. Finally, permission is granted

through the command:

pveum acl modify /vms/<WM_ID> <username>

-role Administrator

Comparative Study of the Support of Academic Clouds Based on Apache CloudStack and Proxmox VE Platforms

351

Table 1: Permissions are required to clone student VMs

from templates.

Permissions Path Propagate

PVEAuditor / +

PVEDatastoreUser /storage +

PVETemplateUser /pool +

PVEVMUser /nodes +

Administrator /vms/VM ID –

Later, using a script, we obtained a list of users

from the LDAP directory and set the appropriate IDs

for them. It is important that when cloning VM, stu-

dents must indicate the ID provided by the teacher for

VM.

3.1 The Academic Cloud Performance

Model

Performance calculations are needed to understand

the number of VMs that can run in the cloud. It is also

needed to understand the computing performance of

these machines, which is required to create their tem-

plates.

Both platforms use CPU and memory redundancy

depending on the number of VMs. VE allows cloud

administrators to do this more ﬂexibly when running

containers instead of VMs.

When creating the service offering templates, we

compared the characteristics of the hardware hosts

(CPU frequency, RAM) with the minimum guest OS

requirements and the number of students. To do this,

we used the inequality:

FRQ = N

∗ F

< FRQ

hosts

where FRQ – the total frequency of VMs processors;

– amount of students; F

– the minimum fre-

quency is recommended for the guest OS; FRQ

hosts

–

total frequency of hardware host processors. The last

value can be found from the ratio:

FRQ

hosts

∑

i=1

where N

– the number of cores in the processor of

the i-th host, F

– CPU frequency of the i-th host.

Unlike Apache Cloudstack, Proxmox VE does not

have performance templates. Appropriate parameters

must be set for each VM during its installation. If a

student clones a VM from a template, he/she cannot

change its performance settings. Instead, when using

the VM, these parameters are available for change.

However, the last two formulas are also valid for the

Proxmox VE platform. In order to compare the com-

puting capabilities of the studied cloud platforms, we

will consider the hardware characteristics of the re-

spective servers (table 2).

Table 2: Main characteristics of academic cloud’s hardware.

N N

MEM

FRQ

hosts

Host0 4 3200 16384 12800

Host1 4 3100 24576 12400

Host2 4 3100 16384 12400

Host3 4 3700 32768 14800

CloudStack 16 90112 52400

Proxmox VE 8 3500 28000 31980

As table 2 shows, the private academic cloud

based on Apache CloudStack has a total frequency

of about 50GHz. And the total amount of memory

is about 90 Gb.

Unfortunately, we do not have the equipment to

deploy identical academic clouds based on the Cloud-

Stack and Proxmox VE platforms. However, we can

assume that the performance of the physical server

with Proxmox VE should be equal to the computing

power of the two servers with CloudStack (Host0 and

Host2). Therefore, in the CloudStack deploymengt,

we disabled Host1 and Host3. During the testing,

Host1 stayed online as a management server in the

CloudStack infrastructure. But it didn’t run any VMs.

Regarding the frequency, two other opposite fac-

tors should be taken into account:

• table 2 shows the base frequency, and processors

can run faster thanks to Turbo-Boost technology;

• hosts run other software (OS, databases, manage-

ment servers, hypervisors). It also consumes re-

sources.

It is well known that the frequency of a modern

processor is not constant. It can increase or decrease

depending on the mode of operation of the CPU.

That’s why we use Processor Base Frequency in the

tables and formulas above. Processor Base Frequency

describes the rate at which the processor’s transistors

are open and close. The KVM hypervisor measures

this processor frequency. It is software for launching

and executing virtual machines in the Apache Cloud-

Stack and Proxmox VE platforms.

Similarly, to determine the required amount of

memory we used the inequality:

MEM = N

∗ MEM

< MEM

hosts

3.2 Veriﬁcation of the Academic Cloud

Performance Model

To check computation performance of both clouds we

loaded the same VMs on two clouds until their CPU

load was 90%.There were several types of test vir-

tual machines (table 3). These types corresponded

to the VM templates that were created for the var-

ious training courses. For example, templates Lin-

AET 2021 - Myroslav I. Zhaldak Symposium on Advances in Educational Technology

352

uxGUI, WindowsWs were used in the course of op-

erating systems, AdvLinux in the course of computer

networks, WindowsSrv in the course of administra-

tion. The Linux OS with CLI worked as Proxmox VE

containers and as regular VMs in CloudStack. We use

the EVE-NG platform based on AdvLinux template

for modelling in the study of computer networks. It

launches its own VMs inside the main Apache Cloud-

Stack VM (Spirin et al., 2019). Such nested virtual-

ization requires more resources.

Table 3: Some templates types of our academic cloud’s per-

formance.

Template F

MEM

FRQ MEM

LinuxCLI 500 500 20 10000 10000

LinuxGUI 1500 2000 20 30000 40000

WindowsWs 1000 2000 20 20000 40000

WindowsSrv 1500 2000 20 30000 40000

AdvLinux 2500 4000 20 50000 80000

The number of VMs that were loaded is shown in ta-

ble 4.

Table 4: Comparison of the number of VMs running on

Apache CloudStack and Proxmox VE platforms.

Template CloudStack Proxmox VE

LinuxCLI* 50 120

LinuxCLI 112 120

LinuxGUI 18 19

WindowsWs 20 22

WindowsSrv 16 16

AdvLinux 12 8

Comparing the data in table 2 and table 3, we can

conclude that our deployments of both platforms pro-

vides about 50 VMs with Linux with command line

interface (CLI), more than 40 VMs with Windows

Workstation, about 35 VMs with Windows Server and

OC Linux with GUI (see table 4). This test showed

that the number of virtual machines loaded on Prox-

mox VE is slightly higher than in Apache CloudStack

(except for the template AdvLinux).

To verify the performance model, we performed

academic cloud testing. The experiment was per-

formed according to the method described in the

article (Algarni et al., 2018). The authors’ model

used traditional benchmarking systems to measure

system performance and individual components like

CPU, memory, cache and disk performance. Like

the previous case, we compared the performance of

two CloudStack hosts and one Proxmox VE. To test

performance of clouds we have created 16 identical

VMs (8 on Apache CloudStack and 8 on Proxmox

VE). They had equivalent 2-core processors, the same

amount of RAM (2 Gb) and the size of disks (14 Gb).

The same Ubuntu server 20.04 x64 was installed on

all these VMs.

We have performed real-tasks tests such as:

• Ab is a tool for benchmarking Apache web server.

Ab was used for measuring of time taken for the

system to respond to a request from a web client.

This can estimate the response efﬁciency of each

VM by measure the number of static web page

requests a server can fulﬁl in one second.

• The LZMA benchmark measures the amount of

time consumed in compressing a ﬁle using its own

compression algorithm.

• John the Ripper is a tool for measure CPU

throughput. In this benchmark, the efﬁciency

of the process management of the hypervisors is

evaluated. This benchmark is a common deci-

pher application which uses diverse ciphers such

as DES and MD5.

• IOzone benchmark is used to evaluate perfor-

mance of ﬁle system. IOzone generates and as-

sesses many ﬁle operations such as read, write and

random read.

Ab was used for measuring of time taken

for the system to respond to a request from a

web client. This can estimate the response efﬁ-

ciency of each VM. Our script have completed

100000 requests with concurrency level 1000

simultaneous requests. The results of all test

available by the URL https://drive.google.com/

ﬁle/d/1m-3oS7cOWyCFUVeBe8t4EnHOcll6ApyX.

Shared table contains the designations of VMC for

VMs Cloudstack and VMP for Proxmox VE ones.

The average values of ab-test are given in the table 5.

Table 5: Ab benchmark results.

Platform Requests per sec Transfer rate, Kb/s

CloudStack 2115,68 95150,89

Proxmox VE 1787,62 80437,84

In the Apache benchmark, CloudStack perfor-

mance is 18% higher in terms of both the number of

requests and transfer rate. A slightly smaller advan-

tage of Apache CloudStack (near 7%) was shown by

the LZMA test. We performed John the Ripper bench-

marks for the case of coding according to DES 128

and MD5 algorithms. Because the test was run on a

single CPU core, the “real” and “virtual” results were

the same. We analyzed the results for the single-salt

case. They indicate (table 6) that Apache CloudStack

has signiﬁcantly advantage under Proxmox VE (38%

and 54% for DES 128 and MD5 algorithms).

Comparative Study of the Support of Academic Clouds Based on Apache CloudStack and Proxmox VE Platforms

353

Table 6: John the Ripper benchmark results

Platform MD5 hash DES 128 hash

CloudStack 14526,13 61055,88

Proxmox VE 10503,75 39524,38

Similar results were obtained in the performance

test of virtual HDDs (ﬁgure 1).

Therefore, it is reasonable to predict that the per-

formance of the virtual machines themselves will be

higher on the Cloudstack platform. It is hoped that

students will feel this in their academic research work.

However, another important factor in this assessment

is the functionality and usability of the interface. The

Proxmox VE console interface is more advanced than

Apache Cloudstack. This fact, together with a good

mobile version of the platform allows you to work

with the cloud from different devices.

The previous test concerned the performance of

the virtual machines themselves. We decided to in-

vestigate the performance of hardware servers (hosts)

with their running hypervisors. This test will allow

us to evaluate the efﬁciency of the use of computing

resources by cloud platforms. To do this we used

the formula to rank hypervisors’ performance pro-

posed by Reddy and Shyamala (Reddy and Shyamala,

2016). According to it, the hypervisor’s score scores

of its CPU and memory.

Score

HSI

= Score

HCPU

+ Score

HMEM

Each of these scores is calculated as the sum of the

CPU and RAM performance of any VM in the cloud.

Score

HCPU

∑

i=1

V Mi

Score

MEM

∑

i=1

V Mi

The CP

V Mi

and MP

V Mi

values are calculated as

follows

V Mi

= α

CPU

ACPU

V Mi

HCP

V Ms

and

V Mi

= α

MEM

V Mi

HMP

V Ms

where α

CPU

and α

MEM

are CPU and memory weight

factors; ACPU

V Mi

and AM

V Mi

are available CPU and

memory resources of VMi; HCP

V Ms

and HMP

V Ms

are

all VMs relative CPU(RAM) utilization with respect

to host.

The formulas for calculating the values HCP

V Ms

and HMP

V Ms

are as follows

HCP

V Ms

TotalCPU −

∑

i=1

ACPU

V Mi

Total

CPU

× 100,

HMP

V Ms

TotalMEM −

∑

i=1

AMEM

V Mi

Total

Memory

× 100,

where ACPU

V Mi

and AMEM

V Mi

are available

CPU(RAM) of VMi.

Alpha coefﬁcients are the ratio of processor re-

sources (memory) of one VM to all VMs. Since

we have all 8 machines the same, for each platform

CPU

and α

MEM

We applied the above formulas and obtained the

following values for the studied platforms (table 7).

Table 7: The obtained scores of Cloudstack and Proxmox

VE platforms.

AB lzma iozone

Score CloudStack 77,24 68,55 349,29

Score Proxmox 79,03 87,18 284,85

Thus, we can state that the score of both plat-

forms based on the Apache benchmark test is almost

the same (the difference is about 1.5%). Proxmox

VE showed the best results in the lzma test (about

27%). The opposite result of the advantage of Apache

Cloudstack is observed in the iozone test. However,

we have doubts about this case, because the disk sub-

systems of the two platforms are signiﬁcantly differ-

ent (Proxmox VE uses a software raid, and Cloud-

stack generally contains two separate physical hosts

with HDDs). Diagrams of the obtained ratings are

shown in the ﬁgure 2.

So, based on the veriﬁcation of the cloud plat-

form performance model, we can say that the Apache

Apache Cloudstack and Proxmox VE have roughly

equal ratings. In naked academic clouds, the perfor-

mance of 2 physical Apache Cloudstack hosts is ap-

proximately equal to the performance of one Proxmox

VE host. In this case, other factors of the cloud plat-

form at the university will be decisive.

4 DESIGNING AND

REALIZATION AN ACADEMIC

CLOUDS’ BACKUP MODEL

The backup model depends on which platform will

be stored. There are general backup principles that

are valid for both platforms. However, the technical

implementation will be different. We had to develop

scripts for the Apache Cloudstack. Proxmox VE has

powerful built-in backup tools.

Experience shows that the task of backup is very

important and time consuming. This is primarily due

to the large amounts of student VMs data in the pri-

vate academic cloud. Large companies develop a dis-

AET 2021 - Myroslav I. Zhaldak Symposium on Advances in Educational Technology

354

Figure 1: Comparison of the results of iozone test.

Figure 2: Comparison of the Apache CloudStack and Proxmox VE scores.

aster recovery plan in this case. But in educational

institutions, IT services work to perform such tasks.

Therefore, they need to develop a model, identify po-

tential risks in the IT infrastructure, consider and im-

plement an appropriate backup system.

The development of a backup strategy requires

the deﬁnition of the main goals and objectives of the

backup, tools and regulations. In general, the problem

of back-up is relevant for almost all IT infrastructures.

When choosing a backup method, the following crite-

ria are important (Angus, 2019):

• backup time to the storage;

• recovery time from backup;

• the number of copies that can be stored;

• risks due to inconsistency of backups, imperfec-

tion of the backup method, complete or partial

loss of backups;

• overhead costs: the level of load on the servers

when performing copying, reducing the speed of

service response, etc;

• the cost of renting all services and storage.

Currently, there are 3 main backup schemes such

as:

• Full. This type of backup creates a complete copy

of all data.

• Incremental. In this case, only ﬁles that have

changed since the previous backup are copied.

The following incremental backup only adds

ﬁles that have been modiﬁed since the previous

backup.

• Differential. The backup program copies each ﬁle

that has changed since the last full backup. Dif-

ferential copying speeds up the recovery process

(Paulsen, 2011).

To save material costs, we use almost no server

equipment and powerful and high-speed network stor-

Comparative Study of the Support of Academic Clouds Based on Apache CloudStack and Proxmox VE Platforms

355

age in our academic cloud installation. Instead, we

decided to use cloud services. For example, the

Google Drive service within the Google Workspace

for Education Fundamentals package offers virtually

unlimited disk space (Oleksiuk et al., 2017). The dis-

advantage of such a repository is the signiﬁcant time

to upload or download backups. This speed will be

limited by the bandwidth of the university’s Internet

channel. The latter requirement can be considered

acceptable, as our implementation of the academic

cloud is used primarily for training rather than for pro-

duction.

To use Google Drive in our own scripts, we need

to use the API of this service. This interface is ac-

cessible through Google Developers Console, a soft-

ware developer service. First you need to create your

own project. Credentials were created to access this

project. We have chosen to access OAuth 2.0 ac-

counts. OAuth is an open authorization standard that

allows a user or application to give and access data

without having to enter a login and password. Ac-

cess tokens are used for this purpose. Each access

token provides access to a speciﬁc client to speciﬁc

resources and for a speciﬁed period of time (Oleksiuk

et al., 2017). After adding a new project, we created

new data for authentication, selected the type of ap-

plication (desktop) and activated the appropriate API

(Google Drive API).

Because templates and ISO images in the Cloud-

Stack infrastructure do not change, but only new ones

are added, we chose the incremental method to back

up the secondary storage. Its implementation was

based on the use of a ready-made utility for synchro-

nizing storage ﬁles. Unfortunately, there is currently

almost no such high-quality utility like Google back-

up and Sync, which is developed for OC Windows.

We analysed several tools such as:

• Gdrive (grive2). Google Drive client with the sup-

port for the new Drive REST API and partial sync.

It can’t provide continuously wait for changes in

ﬁle system or in Google Drive to occur and up-

load.

• Gnome-online-accounts. It is system utility lo-

cated within system’s settings in Gnome GUI. But

it can only be executed in a graphical interface.

• GoSync is a Google Drive client with GUI support

for Linux. It is designed under the GNU General

Public License. The client is not perfect enough,

for example, it has automatic regular synchroniza-

tion every 10 minutes.

• Google-drive-ocamlfuse is a FUSE (Filesystem in

Userspace) ﬁlesystem for Google Drive, written

in OCaml. FUSE is a free module for the kernel

of Unix-like operating systems. It allows devel-

opers to create new types of ﬁle systems available

for users to mount without the root privileges of

Google Drive on Linux (Yallop et al., 2018).

We used the latest utility. Here are its main fea-

tures:

• full read/write access to ordinary ﬁles and folders;

• read-only access to google docs, sheets, and

slides;

• multiple account support;

• duplicate ﬁle handling;

• access to trash;

• storing Unix permissions and ownership;

• support symbolic links;

• streaming through read-ahead buffers.

Some problem was that the utility requires au-

thorization using a browser in a graphical interface.

Therefore, we used an alternative authorization mode.

Since we already had our own OAuth2 client ID and

client secret, we speciﬁed them in the command:

google-drive-ocamlfuse -id

12345678.apps.googleusercontent.com

-secret abcde12345

As the command tries to start the browser on the

server where there is no GUI we formed the necessary

URL as it is written in the documentation on Google

Developers Console. After going to this address, we

received a veriﬁcation code. This code gave access to

folder synchronization to the Google Drive.

For security reasons, we decided to sync not the

Apache Cloudstack secondary storage itself, but a

copy of it from the backup drive. So, we ﬁrst syn-

chronized local folders with the command:

rsync -azvh /export/secondary

/export/sync_secondary/arch_cloud

where /export/secondary – the secondary storage of

Apache CloudStack infrastructure;

/export/sync secondary/arch cloud – the local copy

of this storage.

To synchronize the

/export/sync secondary/arch cloud folder, the follow-

ing command has been added to the server task sched-

uler:

google-drive-ocamlfuse

/export/sync_secondary

It runs every time a server with secondary storage

is loaded.

A backup of all databases is required to restore

the Apache CloudStack cloud infrastructure. These

are such databases:

AET 2021 - Myroslav I. Zhaldak Symposium on Advances in Educational Technology

356

• Cloud. It contains all objects of cloud infrastruc-

ture.

• Cloud usage. A database that contains general-

ized data on resource consumption by the end

user. It is used to obtain statistics and compile

reports.

Since the backup of these databases is quite small,

we decided to store all backups in the cloud stor-

age (Backup Database task, ﬁgure 3). The tradi-

tional database for the Apache CloudStack platform

is MySQL. The main utility for backing up MySQL

databases is mysqldump. Its syntax involves entering

a login name and password. Because the shell script

in Linux is written as a plain-text ﬁle, it will contain

the name of the user’s password (usually the root) of

the database. This is a potential security risk for the

entire server. In order not to leave open the data for

authorization of the database user, we used the “login

path” option. A “login path” is an option group con-

taining options that specify which MySQL server to

connect to and which ac-count to authenticate as. To

create or modify a login path ﬁle, we have used the

mysql conﬁg editor utility. In general, the commands

for creating and archiving a database dump are as fol-

lows:

/usr/bin/mysqldump

--login-path=DailyBackup -u root -A >

$BACKUP_DIR/"_cloud_all_""$date_daily"".sql"

The variable $date daily contains the current date

of the archive. This allows you to see the date of

archiving directly in the ﬁle name.

tar -czf

$BACKUP_DIR/"archive_cloud_all_$date_daily.sql.tgz"

$BACKUP_DIR/"archive_cloud_all$date_daily.sql"

To upload the ﬁles to the server, we used a ready-

made script from GitHub. Here is its launch:

upload.sh "arch_cloud/DB" "$entry" $upl_file

folder_ID "application/x-gzip"

where

• arch cloud/DB – folder for uploading ﬁles;

• $entry – full path to the ﬁle;

• $upl ﬁle – ﬁle name to download;

• folder ID – Google Drive folder ID;

• application/x-gzip – ﬁle MIME-type.

A special refresh token token is required to pro-

vide long-term access of the up-load.sh script to

Google Drive. It can be obtained by curl-calling a

URL such as:

curl -silent "https://accounts.google.com/o/\

oauth2/token" --data "code=<access_token>&

client_id=<client_ID>&

client_secret=<client_secret>&

redirect_uri=urn:ietf:wg:oauth:2.0:oob&

grant_type=authorization_code"

In general, the scheme of backuping of Apache

Cloudstack infrastructure is shown in ﬁgure 3.

Performing backup of primary

repositories (Backup Primary task

(cloud0,cloud1,cloud2,cloud3)) has some difﬁ-

culties. An analysis of Internet sources, management

server databases, and storages ﬁles showed that the

Apache CloudStack platform does not typically use

full copies of disk templates for each VM. This

means that full backups should be made to reduce the

risk of inconsistencies in primary repository archives.

Additionally, it is necessary to prepare a cloud

platform, stopping all VMs. Of course, students need

to form an understanding of the need to turn off their

own VMs. However, in practice this is not always

possible. Therefore, it is necessary to stop all VM

programmatically, by means of a script. This can be

done using the API features of the Apache Cloud-

Stack platform. Using API functions allows the devel-

oper to access data about cloud infrastructure objects.

It is also possible to change the state of these objects.

To generate a query that contains API functions, you

must specify:

• URL of the management server;

• Service construct ”api?”. It contains the path to a

certain API-function, and indicates the beginning

of the parameters that are transmitted using the

GET method.

• Command. It is the name of the API-function.

• ApiKey. The key, that can be generated for each

user account.

• Additional query options separated like GET

queries using the “&” character.

• Response format (JSON or XML).

• Signature of the request.

Regardless of the protocol (HTTP or HTTPS)

used to access the Apache Cloud-Stack API func-

tions, the request must be signed. This allows the

platform to conﬁrm that the request was sent from a

trusted accounting request that has the authority to ex-

ecute the appropriate command. To sign a request, the

developer must have an API key and an account secret

key. They are generated by the platform administrator

(OAuth, 2020).

Here is our bash-script to stop all working users’

VMs

mysql --login-path=DailyBackup -D cloud

-e "SELECT uuid FROM vm_instance WHERE

Comparative Study of the Support of Academic Clouds Based on Apache CloudStack and Proxmox VE Platforms

357

Figure 3: Apache Cloudstack infrastructure backup scheme.

Figure 4: Proxmox VE backup job.

type = \"User\" and state = \"running\";"

> uuid.txt

sed -i ’1d’ uuid.txt

while read LINE; do php -q

cloudstackapi.php "$LINE" ;

done < uuid.txt

In the ﬁrst line we receive in a ﬁle from a database

the list of user VM with a running state. The next

command clears the ﬁrst line because it does not con-

tain a VM. The third line runs the cloudstackapi.php

script. It generates a signature and calls the stopVir-

tualMachine API.

Another way to back up the current state of the

VM is to create their snapshots. The Apache Cloud-

Stack platform provides 2 types of images (Rusyn

et al., 2019):

• VM Snapshot – a hypervisor-driven point-in-time

AET 2021 - Myroslav I. Zhaldak Symposium on Advances in Educational Technology

358

image of a virtual machine’s disks. The exact

mechanism of this is dependent on the hypervisor.

• Volume snapshot – a point-in-time image of a spe-

ciﬁc volume. The process usually involves taking

a VM snapshot and then copying the required vol-

ume to secondary storage and the deleting the VM

snapshot.

This approach requires additional space on the

secondary storage or data coping on the user’s local

disk. Such images can be taken by students from

the web interface of the Apache CloudStack platform.

Performing this action and turning off their own VMs

after the end of their use are important components of

ICT competence of the student.

However, experience shows that not all students

perform these actions. Therefore, these are also worth

automating with scripts. Among the API functions of

the Apache Cloudstack platform are relevant (Sned-

don, 2017).

Another task of backing up our academic cloud

is to estimate the time required to upload data to the

cloud storage. Currently (June 2022) the sizes of our

academic cloud storage is approximately as follows:

• primary250 – 132 Gb;

• primary251 – 70 Gb;

• primary252 – 126 Gb;

• primary253 – 88 Gb;

• secondary – 110 Gb.

Since we make a full copy of the primary storage,

we need to download about 400 GB to the cloud stor-

age each time. Let the speed of the Internet channel

at night be 80 Mbps (10 Mbytes per second). Then it

will take 11 hours to download 400*1024 MB. That’s

a lot. Therefore, we balanced Internet access through

2 providers. At the time of backup, our router routes

hosts cloud0 and cloud1 through the ﬁrst provider,

and cloud2 and cloud3 through the second. In this

case, a full backup takes about 6 hours. This time is

also signiﬁcant, but is acceptable.

Another disadvantage of our scheme is the sig-

niﬁcant time required to download backups from the

Google Drive service. However, this time will be

signiﬁcant if the management or storage servers fail.

This means that we must back up the entire OS of the

management server to fast local area network storage.

As mentioned earlier Proxmox VE provides many

tools, including storage and backup types. To back

up all cloud infrastructure, it is advisable to install

and conﬁgure Proxmox Backup Server. It is an en-

terprise backup solution, for backing up and restor-

ing VMs, containers, and physical hosts. By support-

ing incremental, fully deduplicated backups, Prox-

mox Backup Server signiﬁcantly reduces network

load and saves valuable storage space. Unfortunately,

we do not have the physical resources to install a sep-

arate Proxmox Backup Server. Nevertheless, we have

performed some tasks from the cloud infrastructure

backup model.

Stop Students’ VMs. There is no need to use

Proxmox API to perform this task. We already

have student VM ID (used them to establish access

rights). To stop these VMs on the Proxmox host,

we executed the commands qm shutdown VM_ID or

pct stop container_ID. It is advisable to use such

commands to enable (disable) student VMs and con-

tainers at the beginning (end) of classes.

VM Reservation. We have set up a scheduled

task to back up the cloud infrastructure to an NFS

share. It is connected to our university local network

via an OpenVPN tunnel. This allows us to save the

condition of all machines.

Standard Proxmox VE tools allow cloud admin-

istrator to perform this task ﬂexibly. The following

backup options are available (ﬁgure 4):

• host (node) to be copied;

• storage for which to perform job;

• date and time of the task;

• backup objects such as some VMs, all VMs, all

except some VMs;

• compression method (LZO, GZIP, ZSTD, without

compression)

• copy mode (snapshot, suspend, or turn off VM)

• email notiﬁcation about the status of the job or

errors during its execution

The advantage of this method is the reservation of

working VMs “on the ﬂy”. Like Apache CloudStack

we use VMs backup from Proxmox VE cloud infras-

tructure into Google Drive storage.

5 CONCLUSIONS

The private academic clouds should be used in cloud

based learning environment, as they are necessary for

education of future ICT specialists. Despite the avail-

ability of educational grants from leading cloud ven-

dors, many universities are deploying their own pri-

vate academic clouds. Cloud administrators have a

lot of work to do to maintain and support these aca-

demic clouds. Among these tasks, one of the most

important is to ensure the productivity and elasticity

of the cloud. Solving them will allow them to load the

maximum number of VMs in the cloud infrastructure.

Comparative Study of the Support of Academic Clouds Based on Apache CloudStack and Proxmox VE Platforms

359

In this study, the performance of academic clouds

based on Apache Cloudstack and Proxmox VE was

evaluated. Despite some assumptions in comparison,

we can state that the platforms have approximately the

same performance and therefore can provide the same

amount of VM for students. This is because they both

use the same KVM hypervisor. As our experience

has shown, the process of cloud deployment based on

Proxmox VE is technically simpler and faster. This

is even though we have installed Apache CloudStack

many times and Proxmox VE for the ﬁrst time.

An important task in the maintenance of the aca-

demic private cloud is the backup of its components.

To solve it effectively, you need to use different

backup schemes such as full, incremental, differen-

tial. To save data, it is advisable to use both cloud

and local storage. In any case, administrators should

determine how long it will take to build and restore

the entire cloud infrastructure. It is also advisable to

use the API functions of the cloud platform. This will

automate some maintenance tasks.

To back up Apache Cloudstack, we had to use API

and develop our own tools. In any case, this pro-

cess requires large repositories and fast networks. If

the ﬁrst requirement is not met, it is advisable to use

cloud storage. We did so and copy the backups to

Google Drive. Backup of Proxmox VE infrastructure

is technically simpler. To ensure the full functional-

ity of this process, it is advisable to install a Proxmox

Backup Server. All of the above factors indicate that

it is time to migrate our academic cloud to the Prox-

mox platform. Maybe for a while we’ll use them both

together.

REFERENCES

AL-Mukhtar, M. M. A. and Mardan, A. A. A. (2014).

Performance Evaluation of Private Clouds Eucalyp-

tus versus CloudStack. International Journal of Ad-

vanced Computer Science and Applications, 5(5).

https://doi.org/10.14569/IJACSA.2014.050516.

Algarni, S. A., Ikbal, M. R., Alroobaea, R., Ghiduk, A. S.,

and Nadeem, F. (2018). Performance Evaluation

of Xen, KVM, and Proxmox Hypervisors. Inter-

national Journal of Open Source Software and Pro-

cesses (IJOSSP), 9(2):39–54. https://doi.org/10.4018/

IJOSSP.2018040103.

Amiel, T., ter Haar, E., Vieira, M. S., and Soares,

T. C. (2020). Who Beneﬁts from the Public Good?

How OER Is Contributing to the Private Appropri-

ation of the Educational Commons. In Burgos, D.,

editor, Radical Solutions and Open Science: An

Open Approach to Boost Higher Education, Lec-

ture Notes in Educational Technology, pages 69–89.

Springer Singapore, Singapore. https://doi.org/10.

1007/978-981-15-4276-3

Angus, P. (2019). CloudStack Backup and Recovery

Framework. https://www.slideshare.net/ShapeBlue/

cloudstack-backup-and-recovery-framework.

Bykov, V., Mikulowski, D., Moravcik, O., Svetsky, S., and

Shyshkina, M. (2020). The use of the cloud-based

open learning and research platform for collaboration

in virtual teams. Information Technologies and Learn-

ing Tools, 76(2):304–320. https://doi.org/10.33407/

itlt.v76i2.3706.

Fylaktopoulos, G., Goumas, G., Skolarikis, M., Sotiropou-

los, A., and Maglogiannis, I. (2016). An overview of

platforms for cloud based development. SpringerPlus,

5(1):38. https://doi.org/10.1186/s40064-016-1688-5.

Glazunova, O. (2015). Theoretical and methodological

bases for the design and application of an e-learning

system for future IT specialists in an agrarian univer-

sity. Thesis for the degree of Doctor of Pedagogical

Sciences, specialty 13.00.10 - “Information and Com-

munication Technologies in Education”, Institute of

Information Technologies and Learning Tools of the

NAPS of Ukraine, Kyiv, Ukraine. https://lib.iitta.gov.

ua/9212/.

Glazunova, O. G. and Shyshkina, M. (2018). The Con-

cept, Principles of Design and Implementation of

the University Cloud-based Learning and Research

Environment. In Ermolayev, V., Su

arez-Figueroa,

M. C., Yakovyna, V., Kharchenko, V. S., Kobets,

V., Kravtsov, H., Peschanenko, V. S., Prytula, Y.,

Nikitchenko, M. S., and Spivakovsky, A., editors, Pro-

ceedings of the 14th International Conference on ICT

in Education, Research and Industrial Applications.

Integration, Harmonization and Knowledge Transfer.

Volume II: Workshops, Kyiv, Ukraine, May 14-17,

2018, volume 2104 of CEUR Workshop Proceedings,

pages 332–347. CEUR-WS.org. https://ceur-ws.org/

Vol-2104/paper 158.pdf.

Ilin, D. Y., Volovich, M. E., and Filatov, V. V. (2016).

Analysis of CloudStack Platform Suitability for

Management of Different Cloud Infrastructure

Conﬁgurations. Cloud of Science, 3(3):433–443.

http://web.archive.org/web/20180422194515/http:

//cloudofscience.ru/sites/default/ﬁles/pdf/CoS 3 433.

pdf.

Internet2 (2023). Google Cloud Services (GCP).

https://www.internet2.edu/products-services/

cloud-services-applications/google-cloud-platform/.

Khmelevsky, Y. and Voytenko, V. (2015). Hybrid Cloud

Computing Infrastructure in Academia. In WCCCE

2015 - the 20th Western Canadian Conference on

Computing Education. May 8-9, 2015, Vancouver Is-

land University (VIU), Nanaimo, British Columbia,

Canada. https://doi.org/10.13140/RG.2.1.4082.6647.

Korotun, O. V., Vakaliuk, T. A., and Soloviev, V. N. (2020).

Model of using cloud-based environment in training

databases of future IT specialists. CTE Workshop

Proceedings, 7:281–292. https://doi.org/10.55056/

cte.360.

Lytvynova, S. H. (2018). Cloud-oriented learning environ-

AET 2021 - Myroslav I. Zhaldak Symposium on Advances in Educational Technology

360

ment of secondary school. CTE Workshop Proceed-

ings, 5:7–12. https://doi.org/10.55056/cte.129.

Markova, O. M., Semerikov, S. O., Striuk, A. M., Sha-

latska, H. M., Nechypurenko, P. P., and Tron, V. V.

(2019). Implementation of cloud service models in

training of future information technology specialists.

CTE Workshop Proceedings, 6:499–515. https://doi.

org/10.55056/cte.409.

Microsoft (2023). Student developer resources. https:

//azure.microsoft.com/en-us/resources/students/.

OAuth (2020). Using OAuth 2.0 to Access Google APIs.

https://developers.google.com/identity/protocols/

oauth2.

Oleksiuk, V., Oleksiuk, O., and Berezitskyi, M. (2017).

Planning and Implementation of the Project “Cloud

Services to Each School”. In Ermolayev, V., Bassil-

iades, N., Fill, H., Yakovyna, V., Mayr, H. C.,

Kharchenko, V. S., Peschanenko, V. S., Shyshkina,

M., Nikitchenko, M. S., and Spivakovsky, A., edi-

tors, Proceedings of the 13th International Conference

on ICT in Education, Research and Industrial Appli-

cations. Integration, Harmonization and Knowledge

Transfer, ICTERI 2017, Kyiv, Ukraine, May 15-18,

2017, volume 1844 of CEUR Workshop Proceedings,

pages 372–379. CEUR-WS.org. https://ceur-ws.org/

Vol-1844/10000372.pdf.

Paulsen, K. (2011). 12 - Archives, Backups, and

Linear Tape. In Paulsen, K., editor, Mov-

ing Media Storage Technologies, pages 327–377.

Focal Press, Boston. https://doi.org/10.1016/

B978-0-240-81448-3.00012-5.

Rao, U. H. and Nayak, U. (2014). Data Backups and

Cloud Computing. In The InfoSec Handbook: An

Introduction to Information Security, pages 263–

288. Apress, Berkeley, CA. https://doi.org/10.1007/

978-1-4302-6383-8 13.

Reddy, P. V. V. and Shyamala, K. (2016). New scor-

ing formula to rank hypervisors’ performance com-

plementing with statistical analysis using DOE. Fu-

ture Generation Computer Systems, 61:54–65. https:

//doi.org/10.1016/j.future.2016.02.012.

Rusyn, B., Pohreliuk, L., Vysotska, V., and Osypov, M.

(2019). Method of data dedublication and distribu-

tion in cloud warehouses during data backup. Infor-

mation systems and networks, 6:1–12. https://doi.org/

10.23939/sisn2019.02.001.

Shyshkina, M. (2016). The Hybrid Cloud-based Service

Model of Learning Resources Access and its Evalua-

tion. In Ermolayev, V., Spivakovsky, A., Nikitchenko,

M. S., Ginige, A., Mayr, H. C., Plexousakis, D.,

Zholtkevych, G., Burov, O., Kharchenko, V. S., and

Kobets, V., editors, Proceedings of the 12th Inter-

national Conference on ICT in Education, Research

and Industrial Applications. Integration, Harmoniza-

tion and Knowledge Transfer, Kyiv, Ukraine, June 21-

24, 2016, volume 1614 of CEUR Workshop Proceed-

ings, pages 241–256. CEUR-WS.org.

Sneddon, J. (2017). Mount Your Google Drive

on Linux with google-drive-ocamlfuse.

https://www.omgubuntu.co.uk/2017/04/

mount-google-drive-ocamlfuse-linux.

Spirin, O., Oleksiuk, V., Balyk, N., Lytvynova, S., and

Sydorenko, S. (2019). The blended Methodology

of Learning Computer Networks: Cloud-based Ap-

proach. In Ermolayev, V., Mallet, F., Yakovyna,

V., Kharchenko, V. S., Kobets, V., Kornilowicz, A.,

Kravtsov, H., Nikitchenko, M. S., Semerikov, S., and

Spivakovsky, A., editors, Proceedings of the 15th

International Conference on ICT in Education, Re-

search and Industrial Applications. Integration, Har-

monization and Knowledge Transfer. Volume II: Work-

shops, Kherson, Ukraine, June 12-15, 2019, volume

2393 of CEUR Workshop Proceedings, pages 68–80.

CEUR-WS.org. https://ceur-ws.org/Vol-2393/paper

231.pdf.

Spirin, O., Oleksiuk, V., Oleksiuk, O., and Sydorenko, S.

(2018). The Group Methodology of Using Cloud

Technologies in the Training of Future Computer Sci-

ence Teachers. In Ermolayev, V., Su

arez-Figueroa,

M. C., Yakovyna, V., Kharchenko, V. S., Kobets,

V., Kravtsov, H., Peschanenko, V. S., Prytula, Y.,

Nikitchenko, M. S., and Spivakovsky, A., editors, Pro-

ceedings of the 14th International Conference on ICT

in Education, Research and Industrial Applications.

Integration, Harmonization and Knowledge Transfer.

Volume II: Workshops, Kyiv, Ukraine, May 14-17,

2018, volume 2104 of CEUR Workshop Proceedings,

pages 294–304. CEUR-WS.org. shttp://ceur-ws.org/

Vol-2104/paper

154.pdf.

Spirin, O. M., Vakaliuk, T. A., Ievdokymov, V. V., and

Sydorenko, S. I. (2022). Criteria for selecting a cloud-

based learning management system for a higher ed-

ucation institution. Information Technologies and

Learning Tools, 89(3):105–120. https://doi.org/10.

33407/itlt.v89i3.4958.

Tian, J., Wang, Z., and Li, Z. (2020). Low-cost data par-

titioning and encrypted backup scheme for defend-

ing against co-resident attacks. EURASIP Journal on

Information Security, 2020(1):7. https://doi.org/10.

1186/s13635-020-00110-1.

Vakaliuk, T. (2021). Structural model of a cloud-

based learning environment for bachelors in soft-

ware engineering. Educational Technology Quarterly,

2021(2):257–273. https://doi.org/10.55056/etq.17.

Vakaliuk, T., Spirin, O., and Kontsedailo, V. (2021). For-

mation of digital competence of CS bachelors in the

use of cloud-based learning environments. Educa-

tional Technology Quarterly, 2021(3):388–401. https:

//doi.org/10.55056/etq.26.

Vorozhbyt, A. V. (2017). Creation of multimedia content

of the cloud-based learning environment in technical

lyceum. CTE Workshop Proceedings, 4:59–63. https:

//doi.org/10.55056/cte.325.

Wang, B., Wang, C., Song, Y., Cao, J., Cui, X., and

Zhang, L. (2020). A survey and taxonomy on work-

load scheduling and resource provisioning in hybrid

clouds. Cluster Computing, 28:2809–2834. https:

//doi.org/10.1007/s10586-020-03048-8.

Yallop, J., Sheets, D., and Madhavapeddy, A. (2018). A

modular foreign function interface. Science of Com-

puter Programming, 164:82–97. https://doi.org/10.

1016/j.scico.2017.04.002.

Comparative Study of the Support of Academic Clouds Based on Apache CloudStack and Proxmox VE Platforms

361