Performance Analysis of Cloud Computing Based E-Commerce

Server Using PROXMOX Virtual Environment

Marike Kondoj, Herry Langi, Yoice Putung and Varland Lengkong

Manado State Polytechnic, Indonesia

Keywords: Cloud Computing, Proxmox, E-Commerce, MSMEs.

Abstract: Cloud computing is on-demand service access to computing resources such as networks, servers, storage,

applications, and services. A virtual machine is an efficient and isolated duplicate of an actual device.

Nowadays, virtual machines can simulate hardware even though there is no authentic hardware. So far, the

use of computers as server computers is still one server, one computer, even though the specifications of the

computers used are pretty high and branded. So that many computer resources have not worked optimally to

existing computer specifications.This study aims to build a cloud server for e-commerce services for Micro,

Small, and Medium Enterprises (MSMEs) by optimizing the performance of available computing devices.

The trial was carried out on a laboratory scale by simulating up to 750 users to get the results of the

performance analysis of the installed device. The problems in this study used computer specifications that

support virtual machines. The next step is to create a virtual machine server installed on the proxmox server,

then run it together using existing computer resources. The results of the analysis from these trials show the

server computer with proxmox continues to run well in serving requests for access to 400 – 550 users.

1 INTRODUCTION

Cloud Computing or cloud computing is a

combination of the use of computer technology

('computing') and Internet-based development. Cloud

Storage. Cloud Storage is a metaphor for the Internet,

as storage media are often depicted on computer

network diagrams. Besides being like a cloud in a

computer network diagram, Cloud Storage in cloud

computing is also an abstraction of the complex

infrastructure it hides. Cloud Computing applies a

computing method, namely, capabilities related to

information technology are presented as a service so

that users can access it via the Internet without

knowing what is in it, an expert with it, or have to

control over the technological infrastructure that

helps it (Anggeriana, 2011).

Cloud computing is one of the computing models

that can access anywhere. Cloud computing is on-

demand service access to computing resources such

as networks, servers, storage, applications, and

services (Mell, 2012). Cloud Computing is emerging

and overgrowing, both conceptually and in reality.

Legal/contractual, economics, quality of service,

interoperability, security, and privacy issues still pose

significant challenges (Sen, 2015). A virtual machine

is an efficient and isolated duplicate of an actual

device. At this time, virtual machines can simulate

hardware even though no authentic hardware exists

(Purbo, 2012). So far, the use of computers as server

computers is still one server, one computer, even

though the specifications of the computers used are

pretty high and branded. So that many computer

resources have not worked optimally following

existing computer specifications.

Cloud Computing or cloud computing is a

combination of the use of computer technology

('computing') and Internet-based development. Cloud

Storage. Cloud Storage is a metaphor for the Internet,

as storage media are often depicted on computer

network diagrams. Besides being like a cloud in a

computer network diagram, Cloud Storage in cloud

computing is also an abstraction of the complex

infrastructure it hides. Cloud Computing applies a

computing method, namely, capabilities related to

information technology are presented as a service so

that users can access it via the Internet without

knowing what is in it, an expert with it, or have to

control over the technological infrastructure that

helps it (Anggeriana, 2011).

In a previous study by Zheng et al., experiments

were carried out using AIS data. The results show that

virtualization and Elasticsearch can effectively store

Kondoj, M., Langi, H., Putung, Y. and Lengkong, V.

Performance Analysis of Cloud Computing Based E-Commerce Server Using PROXMOX Virtual Environment.

DOI: 10.5220/0011876000003575

In Proceedings of the 5th International Conference on Applied Science and Technology on Engineering Science (iCAST-ES 2022), pages 741-745

ISBN: 978-989-758-619-4; ISSN: 2975-8246

741

and index Spatio-temporal data with high reliability

and efficiency. This paper is innovative in

implementing cloud virtualization (Yicheng Zheng,

2014). Furthermore, Kovari et al. researched that the

Open Node Virtualization Platform is an open source

CentOS-based server virtualization and management

solution. Proxmox VE is a tweaked Debian

distribution with a specially optimized kernel. The

two platforms are identical but based on different

Linux distributions, with minor differences. This

study analyzed these two open-source IaaS systems

as KVM & OpenVZ-based virtualization solutions

(Kovari, 2012). Research conducted by Min Chen et

al. proposes a new architecture, Affective Interaction

through Wearable Computing and Cloud

Technologies (AIWAC), which includes three

components: collaborative data collection via

wearables, model-enhanced sentiment analysis, and

forecasting, and controlled effective interactions.

Based on the proposed architecture, they present the

AIWAC testbed, design a practical mechanism for

computational-based emotional interaction that can

be subjected to exchange, and discuss its open

problem, which inspires potential research as a new

direction (Chen, 2015). Furthermore, in research

conducted by Suhatman, who designed a centralized

virtualization system using Cloud Computing based

on Virtual Machines capable of building Multi

Servers (DNS, WEB, Mail, and FTP) so that lectures

run smoothly and by the actual client-server network

conditions. This system is applied to 3 Informatics

Engineering Classes with 30 Linux nodes and 30

Windows nodes in each class. In the study,

information on the use of data sources in the form of

memory was up to 94.8%, who only used the CPU as

much as 17.8%, and accessing a Linux server was

much faster than accessing a windows server, namely

the average time to access a Linux server was 1.8

minutes, while to access windows takes an average of

6.20 minutes. So we can conclude that the

performance of Proxmox in terms of resource usage,

especially memory, is very maximal. The research

conducted by Harijanto et al. describes the design and

implementation of digital library software using

Slims digital library with cloud computing using

Proxmox Server for system implementation in a

centralized digital library; this saves infrastructure

and computer server maintenance costs. On the cloud

computing server, which made four digital library

servers with trials running simultaneously, the result

was that the server ran well on the cloud computing

server using Proxmox Server (Harijanto, 2018).

This study aims to build a server by optimizing

the computer's performance using the available

resources on the computer according to its

specifications. The trial in this study uses computer

specifications that support machine virtualization.

The result of this research is to build an e-commerce

service server for coconut-derived SMEs using

virtual machine technology based on cloud

computing using a proxmox server.

The trial in this study builds a cloud server by

optimizing the performance of available computing

devices. The next step is done on a laboratory scale

by simulating up to 750 users to get the version of the

installed device. The results show that the server can

work well for access to 400 - 450 users.

2 METHODS AND SYSTEM

DESIGN

2.1 Proposed Method

The research methodology carried out in the

implementation of this research are:

1. Literature Study

Collect research materials or materials in the form

of; Literature study by collecting various

literature/references related to computer networks

(networking), especially those related to cloud

computing. References used are not only obtained

from libraries (books, scientific journals) but also

from the internet.

2. Analysis, Design, and Implementation

This stage is intended to retrieve, study and

analyze the data obtained from direct

observation/observation on server devices to carry

out the system design process and topology

design. At this stage, an analysis of the server

utility used, server scalability, and execution time

of an application running on a non-virtualized

system is carried out with the same application

running on a virtualized machine in the data

center.

iCAST-ES 2022 - International Conference on Applied Science and Technology on Engineering Science

742

Figure 1: Cloud system design analysis.

The results of the analysis stages described in Figure

1 are used as a basis for considering the

implementation of cloud computing for service

optimization. The flow of the analysis stage is

illustrated in Figure 1.

The system design is illustrated using the

topology shown in Figure 2.

Figure 2: Cloud Computing Topology.

The Proxmox VE server uses version 7.2 with a

server IP of 192.168.100.2; then, a virtual machine is

installed with the Linux Ubuntu Server operating

system version 22.04. On the Linux Ubuntu server,

install Nginx, Mysql, and PHP. A router is set up so

that proxmox and VM Ubuntu servers can access the

internet. Routers are also used so that the public can

access the Ubuntu server/web server using VPN and

port forwarding settings. The modem is used to

provide internet access which is set through the

router. Users can access the web server by accessing

the VPN address and the port forwarded to the web

server.

2.2 System Requirements

At this stage, it will explain who must prepare system

requirements in this study using hardware

specifications, shown in table 1 below.

Table 1: System Requirements Specification.

Name

Specifications

Monitor

14 Inch FHD IPS Anti

Glare

Processor

AMD Ryzen 3 3200U

(2.6GHz up to 3.5GHz)

Graphics

Radeon Vega 3 Graphics

Memory

8GB DDR4

Storage

512GB SSD

2.3 E-Commerce Cloud Server Online

The process of making cloud servers using Proxmox

Virtual Environment, open-source server

management and Linux-based Proxmox VE version

7.2 was used, released on May 4, 2022. Proxmox VE

is installed on a computer with AMD Ryzen 3 3200U

(2.6) processor (GHz up to 3.5GHz), 8GB DDR4

RAM, and 512GB SSD storage media. After

installing Proxmox VE without making any changes,

you get an IP of 192.168.100.2. To access Proxmox

VE requires a computer that is connected and can

communicate with one another. On a computer with

a Windows or Linux operating system installed, we

can create a web browser such as Mozilla Firefox or

Google Chrome and type https://192.168.100.2. If

connected, it will be directed to the login page of

Proxmox VE. After successfully logging in, we can

create a Virtual Machine (VM) that can install various

operating systems. In this experiment, we installed

the Linux Ubuntu Server 22.04 Operating system, and

the installation process was like installing a Linux.

When the installation process is complete, you can

configure the router.

The router in question is setting the router so that

it can connect from the WAN (Internet

Source/Modem) to the virtual machine so that the

virtual machine can connect to the internet. The router

used uses the Mikrotik RouterBoard 750 type, where

Port 1 is a WAN (or connected to a modem) with a

DHCP network source, port two is connected to the

Proxmox VE server, and port three is connected to a

computer that will configure the Proxmox Server and

Virtual Machine.

When successfully setting up the Proxmox VE

router and server and the virtual machine can be

connected to the internet, the software is installed on

the Ubuntu virtual machine. Software installations

Performance Analysis of Cloud Computing Based E-Commerce Server Using PROXMOX Virtual Environment

743

include Nginx as an HTTP server and proxy, MySQL

as a database, and PHP 8 as a compiler to run .php

files.

After installing the software, continue uploading

the MSME E-commerce web file using an FTP Client

and importing a SQL database so that users can access

the E-commerce web by typing the virtual machine IP

on the Mozilla Firefox or Google Chome web

browser so that they can access the E-commerce web.

-commerce and the web server has been successfully

created.

So that the web server can not only be accessed

locally, it is necessary to do a Public IP setting with

VPN Tunnel, which is obtained for free from the

http://web.tunnel.my.id website. After successfully

setting OVPN/VPN Tunnel, the author makes Port

Forwarding settings on the menu Firewall then NAT

where IP and local port forwarding is IP virtual

machine with port 80 as HTTP default port.

3 RESULTS AND DISCUSSION

At this stage, a user trial is carried out, divided into

three parts: the first 250 users, 500 users, and then 500

users.

1. Trial 250 Users Online using the test application

from https://loader.io and access VPN Tunnel

with the address http://id26.tunnel.my.id:11019.

The test was carried out at 07:06 PM, and

production values as in table 2 below.

Table 2: Testing 250 User.

O’clock

07:06 PM

07:07 PM

07:08 PM

Memory

1.13GiB

CPU

0.55%

0.61%

1.59%

Disk

6.69kb

621.23byte

12.51kb

Network

– In

7.27kb

280.71byte

4.11kb

Network -

Out

7.27kb

8.22kb

113.57k

From the test table above, for access to 250 users, one

user is experiencing a timeout, so the real test is only

249 users for a significant difference on the network

side. So the pecifications of the hardware can still

handle 250 users. The results of the analysis using the

application for 250 users can also be seen in Figure 3.

Figure 3: Test results in 250 users using the loader.

2. Trial 500 Users Online using the test application

from https://loader.io and access using VPN

Tunnel with the address

http://id26.tunnel.my.id:11019. The test was

carried out at 08:07 PM, resulting in values in

table 3 below.

Table 3: Testing 500 Users.

O’clock

08:06

08:07

08:08

Memory

1.16GiB

CPU

0.55%

0.56%

1.25%

Disk

0kb

8.55kb

Network

– In

3.77byte

2.72kb

Network

- Out

0byte

78.9kb

From the test table above, for access to 500 users,

there are two users experiencing timeouts, so the real

test is only 498 users, for the significant difference is

on the network side. For the CPU, there is an increase,

but it's around 1.5% from when no one is accessing it.

So for the hardware specifications, it can still handle

500 users. The results of the analysis using the

application for 250 users can also be seen in Figure 4.

Figure 4. Test results in 500 users using a loader.

3. Trial 700 Users Online using the test application

from https://loader.io and access VPN Tunnel

with the address http://id26.tunnel.my.id:11019.

iCAST-ES 2022 - International Conference on Applied Science and Technology on Engineering Science

744

The test was carried out at 5.30 PM resulting in a

value like in table.

Table 4: Testing 750 Users.

O’clock

17:36

17:37

17:38

Memory

1.15GiB

1.16GiB

CPU

1.35%

1.76%

3.53%

Disk

0kb

Network

– In

2.07kb

3.4kb

7.45

Network

- Out

48.42byte

75.7byte

219.1kb

From the test table above, for access to 750 users,

there are 162 users experiencing timeouts, so the real

test is only 400. The rest is not continued because it

has reached its peak, for the significant difference is

on the network side. For CPU, there is an increase,

but it's around 1.5% from when no one is accessing it.

So for the hardware specifications, it can still handle

750 users, but the network has reached its peak.

4 CONCLUSION

Based on the study results, it can be concluded that

the cloud server built for e-commerce services for

Micro, Small, and Medium Enterprises (MSMEs)

using the Proxmox virtual machine can run well for

up to 450 users. Based on the results of trials

conducted for 700 network users, it has reached its

peak, so it is no longer very effective for that many

users.

REFERENCES

A. Kovari and P. Dukan, "KVM & OpenVZ virtualization

based IaaS open source cloud virtualization platforms:

OpenNode, Proxmox VE," 2012 IEEE 10th Jubilee

International Symposium on Intelligent Systems and

Informatics, 2012, pp. 335-339, doi:

10.1109/SISY.2012.6339540.

Anggeriana, Herwin, 2011, Cloud Computing, Jurnal

Teknik Informatika, Vol 1 September 2011.

Chen, M., Zhang, Y., Li, Y., Hassan, M.M. and Alamri, A.,

2015. AIWAC: Affective interaction through wearable

computing and cloud technology. IEEE Wireless

Communications, 22(1), pp.20-27.

Harijanto, Budi., Ariyanto, Yuri.(2015): Desain Dan

Implementasi Digital Library Dengan Teknologi Cloud

Computing Menggunakan Server Proxmox. Seminar

Nasional Komputer dan Informatika Terapan Jurusan

Teknologi Informasi Politeknik Negeri Jember 2015.

http://www.proxmox.com. Diakses pada tanggal 1 Juli

2022.

Leymann, F., 2011. Cloud computing.

Mell, P., Grance, P. (2012):The NIST Definition

ofCloudComputing.http://nvlpubs.nist.gov/nistpubs/Le

gacy/SP/nistspecialpublication800-145.pdf .

Purbo, Ono. W. (2012):Membuat Sendiri CloudComputing

Server Menggunakan Open Source.Yogyakarta: C.V

ANDI OFFSET.

Sen, J., 2015. Security and privacy issues in cloud

computing. In Cloud technology: concepts,

methodologies, tools, and applications (pp. 1585-1630).

IGI global.

Suhatman, Rahmat. "Analisa Performansi Server Cloud

Berbasis Proxmox Ve untuk Multi Server dan Multi

Platform pada Praktikum Administasi Jaringan

Komputer." Jurnal Komputer Terapan, vol. 2, no. 1,

2016, pp. 17-26.

Yicheng Zheng, Feng Deng, Qingmeng Zhu and Yong

Deng, "Cloud storage and search for mass spatio-

temporal data through Proxmox VE and Elasticsearch

cluster," 2014 IEEE 3rd International Conference on

Cloud Computing and Intelligence Systems, 2014, pp.

470-474, doi: 10.1109/CCIS.2014.7175781.

Performance Analysis of Cloud Computing Based E-Commerce Server Using PROXMOX Virtual Environment

745