Enhancing Security in Database Grid Infrastructure for Storage Clusters
N. W. CHANAKA LASANTHA1, RUVAN ABEYSEKARA1, M. W. P. MADURANGA2
1Faculty of Graduate Studies,
IIC University of Technology,
Phnom Penh 121206,
CAMBODIA
2Department of Computer Engineering,
General Sir John Kotelawala Defence University,
SRI LANKA
Abstract: - This research project primarily focused on improving security by addressing vulnerabilities and
creating a cost-scalable storage cluster solution, for applications. The Oracle Grid Infrastructure for Storage
Clusters (OGISC) represents cutting-edge advancements in this field. The research extensively explores the
philosophy of OGISC. Explains its core methodology. Recognizing the weaknesses of existing security
mechanisms, we propose a solution to enhance security measures. A key aspect of our investigation involves
integrating Glusterfs, a system known for its ability to scale linearly. We delve into the architecture of the
solution demystifying the storage scale-out processes to its operation. Our study goes beyond integration by
developing an approach and metadata model specifically tailored for Glusterfs ensuring optimal performance and
robustness. One noteworthy aspect of our research is the application of Glusterfs compression with OpenVPN.
This exploration highlights the benefits derived from this integration emphasizing how OpenVPN enhances
Glusterfs capabilities. Rigorous analysis stages serve as the foundation for our findings resulting in a forward-
thinking solution. Finally, we conclude this research with an examination of avenues for future exploration in
this dynamic field.
Key-Words: - Security Enhancement, Cost-Scalable Storage Cluster, Oracle Grid Infrastructure for Storage
Clusters (OGISC), Glusterfs Integration, Storage Scale-Out Processes, OpenVPN and Glusterfs
Compression.
Received: July 12, 2022. Revised: August 27, 2023. Accepted: October 8, 2023. Published: November 30, 2023.
1 Introduction
This project has been extended and generated
important benefits for the various stakeholders by
enhancing the security availability of the cluster
system in secured Oracle grid infrastructure storage
cluster (OGISC) delivery as well as evaluation.
OGISC was heavily focused on security weaknesses
and cryptographic backdoors over the distributed
storage structure with high availability at a lower
cost than expensive alternatives. The OGISC
strongly addressed the specific major problems by
facilitating adaptive software and hardware
resources, that make it very efficient and, also
provide usage capacity on demand over the securing
mechanism, [1]. In addition, it is very low-cost
effective compared to the traditional system, which
effectively balances the load, sessions, and security
weakness over the OGISC concept. The proposed
architectural solution operates over different
geographically clustered databases including the
secure sheared concept, [2].
2 The Philosophy of OGISC
2.1 Hypothesis
1. The SSL/TLS ETE encryption at which storage
cluster can perform secured tunnel security and
accessibility among server nodes against
compromising situations over the shearing storage
area by using encryption algorithms such as RSA,
AES were incorporated while the HMAC function
WSEAS TRANSACTIONS on COMPUTERS
DOI: 10.37394/23205.2023.22.27
N. W. Chanaka Lasantha, Ruvan Abeysekara, M. W. P. Maduranga
E-ISSN: 2224-2872
233
Volume 22, 2023
makes use of a hashing algorithm were built in
OpenVPN.
2. The Initial key exchange (IKE) mechanism of the
SSL/TLS process can be tightly coupled with the
HMAC handshake method over the DH and STC
key exchange algorithms. Also, a static key among
both peer nodes before the certain tunnel was started
of OpenVPN architecture to mitigate risk by
adhering to The X.509 global standard according to
the formation of the public key certificate with
periodic key origin renegotiation process.
3. The digital signature and certificate make the
concept of fingerprinting security mechanism which
can digitally sign by genuine message sender by the
trusted originator of the content. Also, the one-way
hash function can check the integrity of the message.
The multifactor authentications of OpenVPN have
strongly mitigated the account hijack situation. It can
have a powerful firewall with a manageable routing
framework.
4. The intelligent self-driven metadata algorithm of
the GlusterFS has mainly focused on the brick server
to relocate files which can provide flexibility to add
and delete VMs by continuing operations instead of
traditional systems. It provides large scale-out
architecture and minimum overhead on the cluster
including the ability of data performance.
5. OpenVPN can engage highly compressed data
packets to improve performance without adding
extra overhead to OpenVPN protocol that is
inherited from cryptographic functionality using
AES-256-GCM. It contains a pushed routing feature
among TCP or UDP traffic with fail-safe
functionality.
6. The mechanism of the SPARC-M8 processor has
hardware-assisted encryption data which can tightly
couple with Linux kernel while Oracle ZFS storage
appliances encrypt only at the file system level.
Also, it does not provide compressions, replication,
reduplication direct NFS support while direct NFS
client optimizes NFS operations.
2.2 Methodology
Fig. 1: Overview of Methodology.
Figure 1 shows that it is necessary to follow a
certain methodology or scientific approach to
conduct research. During the literature review, a
research gap was found in the case of the security
impact of secured OGISC in ABC LTD. and it was
identified as the research problem to be addressed in
this study.
2.3 Concept of Proposed Architectural
Solution
Secured OGISC service facilitation has been
achieved in more remote locations as well as in
urban backend areas to access the Oracle database at
any time without zero downtime under heavy
encryption algorithms support with ultimate data
compression mechanism to provide maximum
performance between peers’ nodes.
Fig. 2: The Proposed Network Architecture.
Figure 2 shows that encrypted connectivity
tightly binds with the GlusterFS nodes before the
WSEAS TRANSACTIONS on COMPUTERS
DOI: 10.37394/23205.2023.22.27
N. W. Chanaka Lasantha, Ruvan Abeysekara, M. W. P. Maduranga
E-ISSN: 2224-2872
234
Volume 22, 2023
mounting process to provide secure connectivity for
the strong reaction against cybercrime mitigation
situations successfully instead of open connectivity
among the storage cluster.
2.4 Existing Security Weaknesses of
Mechanism
Fig. 3: Common CPU vs Oracle S7 CPU.
Figure 3 shows that the Operation of Oracle
SPARC Processor has an advanced encryption
capability for threat mitigation while the database
instances faced a huge prime number de-
factorization attack force by cryptanalyst. The
SPARC servers used their dedicated contribution of
hardware-assisted encryption security devices, [3].
Oracle has developed a powerful and identical
high-performance processor that was the target of
security against known threats by hackers, in
addition, Oracle created the powerful CPU called
SPARC M7. It was made up of traditional 32 CPU
cores to extend 512 CPU cores successfully.
Fig. 4: SQL in Silicon.
According to Figure 4, The 4.1GHz 32 cores and
256 threads CPU is touted to focus on the highly
demanding workload flow with the specific
enhanced high-performance architecture, the
SPARC M7 processor has incorporated advanced
software techniques and not only focused on
increasing the performance of the CPU with major
improvements to mitigate against programming
errors were caused to happened serious security
breaches, [4].
Fig. 5: Silicon Secured Memory.
Based on Figure 5, The Silicon Secured Memory
structure of the SPARC M8 processor has its
powerful encryption engine associated with it, and
stronger encryptions and hashing algorithms are
included such as RSA, 3DES, SHA 256, SHA 512,
DH, MD5, and ECC. Therefore, to protect physical
files from security threats and CPU core’s scalability
in the dynamic way of predictable optimization to
overcome the slowness of the encryption process
under heavy load, [5].
Fig. 6: Transparent Data Encryption.
Based on Figure 6, the execution of certain
machine code was very straightforward, and it
became very complex. At the same time, poorly
programmed code examples have been caused by
pointers attempting to get access to an allocated
loaded memory location. This behavior was a
maliciously exploited situation called the buffer
overflow (over-read attack). The attacker took
advantage of the security weakness of the code
section in such a buffer overflow attack. Also, this
badly allowed the attacker to modify (write) to an
adjacent physical memory location in the flow of
buffer overflow attack.
The above condition of the attack obtained data
segments of the memory due to the case of a random
overread attack. The side effect of the possibility of
altering the ability of a program and executing
attackable malicious code returned as informative
details and access to exploited by an attacker, else it
breaches system security, [6].
The CVE-2018-3639 vulnerability was the side
channel disclosure attack to expose confidential
information. All of them were identified in the same
WSEAS TRANSACTIONS on COMPUTERS
DOI: 10.37394/23205.2023.22.27
N. W. Chanaka Lasantha, Ruvan Abeysekara, M. W. P. Maduranga
E-ISSN: 2224-2872
235
Volume 22, 2023
category of attacks. but it was different from the
formation of remote code execution was exploited.
In addition, these are the attacks that never allow for
an unauthorized party to obtain access to a machine.
However, it would have allowed a certain external
party to access confidential unauthorized data, [7].
The Solaris-10 SPARC allowed a third-party
unauthenticated cyber-attacker with the ICMP-
enabled platform to compromise the SPARC system.
After the successful attempts of attacks of this
identified vulnerability was enabled backdoor access
as the unauthorized ability to crash the system with
DOS attacks over the SPARC systems successfully.
Furthermore, the above exploitable vulnerability
allowed deliberative outside privileges attackers to
log in to the operation infrastructure with remote
shell execution privileges, [8].
2.5 Linear Scaling and Introducing
GlusterFS
The GlsuterFS Striped volume successfully except
for strips that can distribute over a very large
number of bricks, additionally, the numbers of
bricks must be of the multiplication of the numbers
of bricks, were led to increased volume size. The
method of linear scaling was the much cuticle phase
within the traditional cluster storage field.
Also, when an organization needs to increase
performance by twice, that there is clustered storage
system must deliver twice the performance and
throughput within the same average response of time
gap per external clustered file system. Equally, if
they wanted to increase either capacity without
decreasing performance or having a non-linear
return in capacity. Unfortunately, most storage
clustering systems do not perform linear scaling.
Simply, when an organization needs to double the
disk size of the available storage pool, then it must
be required to provide enough peak CPU processing
power, [9].
The latency takes place as the result of responses
across the storage cluster network connected over
the distributed cluster server nodes in those
traditional types of storage system architectures and
recently always impacted the overall performance.
The overhead was the major risk while each node in
an unacceptable situation led to risk.
Also, this was one of the main reasons for linear
scalability caused by the reduced performance of
traditional storage distributed architectures. Most of
the traditional storage systems demonstrated
logarithmic scalability when capacity grows very
slowly as it gets larger. This was due to the rapidly
increased average overhead required to maintain
data flexibility, [10].
2.6 The Ultimate GlusterFS Architecture
When a traditional storage system adds more and
more files and more server nodes with more disk
arrays, then the centralized metadata server becomes
the performance chokepoint, [11].
The GlsuterFS had the mechanism to find a file
algorithmically. Therefore, all GlsuterFS storage
server nodes within the cluster have a specially
developed intelligent algorithm to locate any piece
of physical data without depending on the metadata
in a separate server instead of a traditional system.
The Hashing Algorithm was used when scaling out a
storage system, data, and workload. when the
storage nodes have been physically located in many
different locations as independent storage and cluster
nodes to resolve difficulties while retrieving files
and locations, [12].
2.7 Storage Scale-Out Process of the
GlusterFS
The GlusterFS was designed to successfully provide
extended scalable architecture on both capacity and
performance with minimized overhead problems.
This illustrated that the storage cluster system must
be able to scale down or scale up among multiple
dimensions. Also, by aggregating the CPU, HDD
arrays, and I/O buses of very large numbers of a
low-cost system without expending lots of money on
expensive resources. In general, an enterprise
organization must be able to implement a very
scalable and performant cluster storage pool.
Fig. 7: Logarithmic Vs Linear Improvement.
Figure 7 shows that Logarithmic vs. linear
Improvement over the GlusterFS has a unique and
WSEAS TRANSACTIONS on COMPUTERS
DOI: 10.37394/23205.2023.22.27
N. W. Chanaka Lasantha, Ruvan Abeysekara, M. W. P. Maduranga
E-ISSN: 2224-2872
236
Volume 22, 2023
advanced architecture designed to deliver huge
benefits for expanding scalability. Also, it can be
defined as, more units for more capacity, more CPUs
as well as more I/O capability which was archived
over the storage cluster based on GlusterFS while
avoiding the system overhead and, the critical risk
associated when it has very large numbers of server
nodes in the synchronization process.
Practically, both performance and the capacity
of the storage cluster must need scaling out in the
manner linearly in the GlusterFS architecture. The
Illustration of GlusterFS cluster storage scalability,
Figure 8, below shows how the enhancements of
both performance and capacity have been archived
over the baseline system. As an example, if they
expected to obtain both four times capacity and
performance, they must be distributed among 8
servers, [13].
Fig. 8: Linear Scaling of GlsuterFS.
Figure 8 shows the illustration of theoretical
numbers only for the example purpose of
demonstrating and actual storage performance has
been tested to prove the credibility of the linear
scaling. The resultant of the storage cluster linearity
has been demonstrated in Figure 9 below that shows
the data write process of throughput scaling linearly
started from 100MB/s to 800MS/s among the eight
servers using the 1 Gigabit Interface environment.
Fig. 9: Cryptographic Operations with OpenSSL.
Therefore, the GlsuterFS storage cluster has
been successfully deployed in the massive scale-out
concept in practice. As a result, it can successfully
deploy in the petabyte size clustering solution, [14].
2.8 The Algorithmic Approach and Metadata
Model of GlusterFS
The metadata separate location was the main single
point of failure, performance overhead as well and
reliability concerns of most distributed cluster-based
storages. There was no separate metadata server for
the information data because the location can
determine independently when the other nodes are
up or down.
Therefore, GlusterFS called the above
algorithmic file location mechanism called Elastic
Hashing Algorithm, and it benefited had unique
advantages of the GlusterFS architecture, [15].
Fig. 10: Node adds and deletes effect mitigation.
Figure 10 shows that the GlusterFS elastic
hashing algorithm is based on a special mechanism
called the Davies-Meyer hashing algorithm. The
GlsuterFS way of algorithmic approach was unique
in any clustered directory tree and ran through the
GlusterFS elastic hashing algorithm. In the real
infrastructure environment, if the array of disks fails,
the capacity of the cluster is used up, and files need
to be redistributed over the cluster to get back into
the smooth working state without interrupting saved
data, [16].
2.9 GlusterFS Compression over OpenVPN
The dedicated compression translator mechanism
was embedded into the GlsuterFS architecture to
archive high data compression and decompression
processes while transferring between clients and
bricks over Cluster nodes. The bricks of the server
nodes compress the data before transferring it to the
client side. The overall throughput measurement has
been calculated using the open-source tool called
WSEAS TRANSACTIONS on COMPUTERS
DOI: 10.37394/23205.2023.22.27
N. W. Chanaka Lasantha, Ruvan Abeysekara, M. W. P. Maduranga
E-ISSN: 2224-2872
237
Volume 22, 2023
iperf after turning off all pre-configured
authentication and encryption operational processes
as the cipher none state, and then only triggering the
iperf connectivity test.
As a result, the way of compression comparison
was illustrated actuality over the existing GlsuterFS
storage cluster. The process of decompression and
compression operations was done by using the
developed ZLIB library bundle. In addition, the
enhanced speed of the compressed data, [17].
3 A. The SSL/TLS Connectivity
over the GlusterFS
Figure 11 shows, that there was a set of OpenSSL
speed test commands to show that the decryption
and encryption absolute performance over different
networks depends on identical hardware units upon
the encryption key already used. By default, the
OpenVPN packets were configured at 1500 bytes.
Also, The BlowFish cipher has been divided by
CPU clock speed and its performance is hardly
bound purely by CPU clock rate. But Older types of
CPUs operating at a higher clock plus speed, cause
them to outperform compared with newer CPUs.
Fig. 11: Speed test of AES-256-GCM.
In general, an OpenVPN server must handle
many numbers of VPN client connections, then
because of that cryptographic cipher was a proper
choice, [18].
B. Benefit and OpenVPN Mechanism over
GlusterFS.
The OpenVPN was running over TCP by using port
number 443, so the bunch of traffic was
distinguishable from the general way of TLS
traffic. The Deep Packet Inspection process can be
used to filter out OpenVPN traffic.
In addition, the major difference between
Browser-based TLS and OpenVPN TLS was the
way packets were signed. OpenVPN has been
offered to mitigate security attacks from DoS attacks
by using special signing packets using the method
over the control channel by using a static key
generated randomly called TLS auth Key, [19].
A type of hardware token and smart cards were
typically very small devices that could be embedded
into a chip. Also, it was highly responsible for
securely generating, storing, and managing SSL-
based private keys. Additionally, that feature was
validated certificates and private key pairs were
securely stored in a portable single device, [20].
C. Sampling Calculation and Process
Table 1. Morgan Table
Table 1 shows that the questions have been
asked to obtain information about the occupational
pattern, potential members of the team, critical
assets with priority under the area of concern, and
identification, and security requirements on critical
assets, [21].
D. Conceptual Framework
Fig. 12: The Conceptual Framework.
Figure 12 shows the theoretical framework that
structures the sections of the study that need to be
WSEAS TRANSACTIONS on COMPUTERS
DOI: 10.37394/23205.2023.22.27
N. W. Chanaka Lasantha, Ruvan Abeysekara, M. W. P. Maduranga
E-ISSN: 2224-2872
238
Volume 22, 2023
covered and it can help to determine the problem
area.
Research questions have been addressed and
methodology to find the solution for this problem.
E. Analysis Stages.
Fig. 13: Critical Value Identification.
Figure 13 shows that the CSQ was used to
effectively test hypotheses regarding the distribution
of certain observations between different categories,
[22].
Fig. 14: The Illustration of Correlation Coefficient.
Figure 14 shows that the Correlation Analysis
method was developed to properly assess the direct
and indirect relationship between variables. Also, the
affected side of the Correlation analysis process has
been mentioned as the strength of the correlation,
[23].
F. NMAP Scan over GlusterFS Nodes
Illustrates the resultant of the NMAP scan using the
techniques of TCP SYN scan. It was initiated using
different scanning principles such as ARP Ping Scan,
SYN Stealth Scan, Service Scan, and RPC Bind
Scan against each node among Storage GlusterFS.
The Next step was to attack both Oracle RAC
and GlsuterFS Cluster using the Armitage
Application has been powered by Metasploit
Framework as an automated attack method that was
already generated by Pen Test. It failed to exploit the
required server nodes by using known and
discovered mid-range level vulnerabilities due to a
recent Linux OS update that was made over the
cluster. As a result of the overall Pen Test, it failed
to attack any single Node of the Server due to the
patch management process with the latest OS
Update and Upgrade of the rest of the Application
behind the cluster node in the Oracle RAC.
4 Conclusion
Fig. 15: Statistical, Octave, and Pen test integration.
Figure 15 shows that the validation process was
illustrated in the six hypothesis statements due to
various affected sizes among the independent
variable according to the dependent variable. The X2
> 0 conditions have proved the validation of the
objects which was involved in the data collection
and analysis phase.
The output has been directly inherited into the
Octave Framework for the further technical level of
vulnerability assessment with organizational risk
factors in the particle way and a proven good level
of security strength regarding the provided solution.
After the OS, Kernel, and application update
process, it was reduced up to 98%. Also, the
organizational vulnerabilities were reduced with the
Security Policy of ABC LTD after the embedded
with Octave stage with existing ABC LTD’s security
policy as well.
The penetration test occurred to attack the IT
assets after an anonymous way of the test method
and got the sufficient level of the good security level
of the provided solution with the above stages.
WSEAS TRANSACTIONS on COMPUTERS
DOI: 10.37394/23205.2023.22.27
N. W. Chanaka Lasantha, Ruvan Abeysekara, M. W. P. Maduranga
E-ISSN: 2224-2872
239
Volume 22, 2023
4.1 Proposed Solution and Future Research
Fig. 16: The Solution with Enhancements.
Figure 16 shows that must be very adaptive and
important to propose a system for the monitoring
purpose of storage volumes when ABC LTD plans
to establish capacity planning or performance tuning
scheduled activities over the GlusterFS OGISC.
Also, that can monitor Gluster storage volumes
using different customizable parameters and can
reuse those outputs to identify and clear
understanding of troubleshooting issues. The Zabbix
monitoring system could enable information
gathering such as CPU, physical memory, Raid disk,
Network Interfaces, Swap memory, cluster, cluster
volume, brick over nodes, LVM host, Glusterd
quota, geo-replication, self-heal so on using https
secured management web interface.
However, the proactive analysis mechanism can
enable detection processes and events that are not
specifically highlighted by an automated IDS system
and it should cause to use to limit the broader range
of impact as well as the cost of an incident.
The requirement of correlation of logs and
events over a variety of devices was a very critical
area of secured storage cluster over Oracle grid
incident response activity such that assists an
organization in the assessment process over the
possibility of the impact of a network compromise.
Also, that must be informed toward the Security
team of ABC LTD to necessary for mitigation. The
security hardware or software-based product should
have identified when an intruder randomly
performed an anonymous remote attack against ABC
LTD’s OpenVPN and GlsuterFS applications, but
sometimes it does not have the alert rules to identify
the attacker’s random penetration through the ABC
LTD’s network from the initial stage of compromise.
By correlating the original alert event with log files
from the GlusterFS storage cluster, OpenVPN,
Syslog, access log, Oracle database, and
authentication servers, as well as events, were
generated on the server nodes, greatened visibility of
the extent of the compromise which must be
established.
References:
[1] Y. Wei, R. Ye and X. Chen, "Oracle RAC
performance analysis on VMware Virtual
SAN," 2019 IEEE/ACIS 18th International
Conference on Computer and Information
Science (ICIS), Beijing, China, 2019, pp. 562-
565, doi: 10.1109/ICIS46139.2019.8940347.
[2] Y. Gong, Y. Wu, K. Pan, J. Chang and J. Xu,
"Research and application of distributed
storage technology in power grid enterprise
database," 2019 IEEE Innovative Smart Grid
Technologies - Asia (ISGT Asia), Chengdu,
China, 2019, pp. 1706-1710, doi:
10.1109/ISGT-Asia.2019.8881152.
[3] Oracle LTD, "Oracle’s SPARC M7
Processor–Based Servers," oracle.com,
[Online].
http://www.oracle.com/technetwork/server-
storage/hardware-solutions/oos-for-secure-
oracle-database-2736047.pdf (Accessed Date:
November 22, 2023).
[4] Oracle LTD, "Enhanced data centre security
with Oracle SPARC and Oracle Solaris,"
oracle.com, [Online].
http://www.oracle.com/us/products/servers-
storage/sparc-your-power-wp-3033447.pdf
(Accessed Date: November 22, 2023).
[5] S. Pendse, V. Krishnaswamy, K. Kulkarni, Y.
Li, T. Lahiri, V. Raja, J. Zheng, M. Girkar, A.
Kulkarni, "Oracle Database In-Memory on
Active Data Guard: Real-time Analytics on a
Standby Database," 2020 IEEE 36th
International Conference on Data
Engineering (ICDE), Dallas, TX, USA, 2020,
pp. 1570-1578, doi:
10.1109/ICDE48307.2020.00139.
[6] J. Stuecheli, S. Willenborg and W. Starke,
"IBM’s Next Generation POWER Processor,"
2019 IEEE Hot Chips 31 Symposium (HCS),
Cupertino, CA, USA, 2019, pp. 1-19, doi:
10.1109/HOTCHIPS.2019.8875663.
[7] D. Chen, G. Jacques-Silva, Z. Kalbarczyk, R.
K. Iyer and B. Mealey, "Error Behavior
WSEAS TRANSACTIONS on COMPUTERS
DOI: 10.37394/23205.2023.22.27
N. W. Chanaka Lasantha, Ruvan Abeysekara, M. W. P. Maduranga
E-ISSN: 2224-2872
240
Volume 22, 2023
Comparison of Multiple Computing Systems:
A Case Study Using Linux on Pentium,
Solaris on SPARC, and AIX on POWER,"
2008 14th IEEE Pacific Rim International
Symposium on Dependable Computing,
Taipei, Taiwan, 2008, pp. 339-346, doi:
10.1109/PRDC.2008.35.
[8] S. Menaka, K. Navaneethakrishna, J. A.
Blesswin Geo Sam, V. K. Hariharasuthan, S.
Murugesan, N. Bharathiraja, "Cyber Security
Tool for Combat Remote Work
Vulnerability," 2023 Fifth International
Conference on Electrical, Computer and
Communication Technologies (ICECCT),
Erode, India, 2023, pp.1-4, doi:
10.1109/ICECCT56650.2023.10179806.
[9] Gluster Community, "Distributed Striped
Volume," glusterdocs-beta.readthedocs.io,
[Online]. http://glusterdocs-
beta.readthedocs.io/en/latest/overview-
concepts/volume_types.html (Accessed Date:
November 22, 2023).
[10] S. Y. Pan, T. Morris, and U. Adhikari,
Developing a hybrid intrusion detection
system using data mining for power systems,
IEEE Trans. Smart Grid, vol. 6, no. 6, pp.
3104–3113, 2015.
[11] Gluster Inc, "Distributed Striped Glusterfs
Volume," docs.gluster.org, para, [Online].
https://docs.gluster.org/en/v3/Quick-Start-
Guide/Architecture/. (Accessed Date:
November 22, 2023).
[12] Azure Inc, "Gluster storage architecture"
azure.microsoft.com, [Online].
https://azure.microsoft.com/mediahandler/files
/resourcefiles/771b82cd-8b2a-4466-886d-
43abf80b14c2/Implement_GlusterFS_on_Azu
re.pdf (Accessed Date: November 22, 2023)
[13] Gluster Inc, "DHT (Distributed Hash Table)
Translator," docs.gluster.org, [Online].
https://docs.gluster.org/en/v3/Quick-Start-
Guide/Architecture/ (Accessed Date:
November 22, 2023).
[14] RedHat Inc, "Workload-Optimized Distributed
File System Clusters., "redhat.com, [Online].
https://www.redhat.com/cms/managed-files/st-
RHGS-QCT-config-size-guide-technology-
detail-INC0436676-201608-en.pdf (Accessed
Date: November 22, 2023).
[15] Sudarshan D. "Simple Application of
GlusterFs: Distributed file system for
Academics", International Journal of
Computer Science and Information
Technologies, Vol. 6 (3), 2015, 2972-2974.
[16] RedHat Inc, "No Metadata with The Elastic
Hashing Algorithm., "access.redhat.com,
[Online]
https://access.redhat.com/documentation/en-
us/red_hat_gluster_storage/3.1/html/administr
ation_guide/no_metadata_with_the_elastic_ha
shing_algorithm (Accessed Date: November
22, 2023).
[17] Gluster Inc, "On-Wire Compression +
Decompression, "staged-gluster-
docs.readthedocs.io, [Online]. https://staged-
gluster-
docs.readthedocs.io/en/release3.7.0beta1/Deve
loper-guide/network_compression/ (Accessed
Date: November 22, 2023).
[18] Gluster Community, "Setting up GlusterFS
with SSL/TLS," docs.gluster.org, [Online].
https://docs.gluster.org/en/latest/Administrator
%20Guide/SSL/#setting-up-glusterfs-with-
ssltls (Accessed Date: November 22, 2023).
[19] Packt LTD, "OpenVPN Cookbook.
Birmingham”, UK.: Packt Publishing Ltd.,
2017, pp. 279-282.
[20] Packt LTD, "OpenVPN Cookbook.
Birmingham”, UK.: Packt Publishing Ltd.,
2017, pp. 19-90.
[21] MaCorr Research Inc, "Sample Size
Terminology., "macorr.com, [Online].
http://www.macorr.com/sample-size-
methodology.htm (Accessed Date: November
22, 2023).
[22] S. Tsumoto and S. Hirano, "Degree of
freedom and numbers of subdeterminants in
contingency table," 2012 IEEE International
Conference on Granular Computing,
Hangzhou, China, 2012, pp. 481-486, doi:
10.1109/GrC.2012.6468603.
[23] DJS Research Inc, "Sample Size
Terminology., "djsresearch.co.uk, [Online].
https://www.djsresearch.co.uk/glossary/item/c
orrelation-analysis-market-research.
(Accessed Date: November 22, 2023).
WSEAS TRANSACTIONS on COMPUTERS
DOI: 10.37394/23205.2023.22.27
N. W. Chanaka Lasantha, Ruvan Abeysekara, M. W. P. Maduranga
E-ISSN: 2224-2872
241
Volume 22, 2023
Contribution of Individual Authors to the
Creation of a Scientific Article (Ghostwriting
Policy)
The authors equally contributed in the present
research, at all stages from the formulation of the
problem to the final findings and solution.
Sources of Funding for Research Presented in a
Scientific Article or Scientific Article Itself
No funding was received for conducting this study.
Conflict of Interest
The authors have no conflicts of interest to declare.
Creative Commons Attribution License 4.0
(Attribution 4.0 International, CC BY 4.0)
This article is published under the terms of the
Creative Commons Attribution License 4.0
https://creativecommons.org/licenses/by/4.0/deed.en
_US
WSEAS TRANSACTIONS on COMPUTERS
DOI: 10.37394/23205.2023.22.27
N. W. Chanaka Lasantha, Ruvan Abeysekara, M. W. P. Maduranga
E-ISSN: 2224-2872
242
Volume 22, 2023
