Web Server Performance Improvement Using
Web Server Performance Improvement Using
Authors’ contributions
This work was carried out in collaboration among all authors. All authors read and approved the final
manuscript.
Article Information
DOI: 10.9734/AJRCOS/2021/v10i130234
Editor(s):
(1) Dr. Dariusz Jacek Jakóbczak, Koszalin University of Technology, Poland.
Reviewers:
(1) Bikramaditya Das, VSSUT, India.
(2) N. K Choudhari, Priyadarshini Bhagwati College of Engineering, India.
Complete Peer review History: http://www.sdiarticle4.com/review-history/70184
ABSTRACT
Today, web services rapidly increased and are accessed by many users, leading to massive traffic
on the Internet. Hence, the web server suffers from this problem, and it becomes challenging to
manage the total traffic with growing users. It will be overloaded and show response time and
bottleneck, so this massive traffic must be shared among several servers. Therefore, the load
balancing technologies and server clusters are potent methods for dealing with server bottlenecks.
Load balancing techniques distribute the load among servers in the cluster so that it balances all
web servers. The motivation of this paper is to give an overview of the several load balancing
techniques used to enhance the efficiency of web servers in terms of response time, throughput,
and resource utilization. Different algorithms are addressed by researchers and get good results
like the pending job, and IP hash algorithms achieve better performance.
Keywords: Web server; load balance; cluster; NginX server; apache sXerver.
_____________________________________________________________________________________________________
48
Ibrahim et al.; AJRCOS, 10(1): 47-62, 2021; Article no.AJRCOS.70184
challenging problems to solve in recent years The use of a dynamic architecture for cluster
[39]. self-reconfiguration is one solution to this
challenge [51].
As the amount and hours of movies available
within a company grows, so does demand, By replicating material, data-sharing scientific
necessitating rapid video streaming services [40]. communities use storage systems as distributed
Cloud-based services are not cost-effective and data stores [52]. The same dataset can exist in
are not a viable solution for storing the growing numerous locations in such highly duplicated
amount of video data often saved and used only settings and can thus be retrieved from any of
within a single organization, such as a university them [53]. Because the datasets of interest are
[41]. typically massive, increasing download speeds
through server selection or co-allocation can
For clustered web applications, dynamic provide significant benefits [54].
application placement has a significant impact on
system performance and user experience [42]. Massive Multiuser Virtual Environments
Existing approaches claim to maximize (MMVEs) are currently attracting a lot of interest.
throughput, balance resource use among Beyond the commotion, some exciting notions
servers, and reduce the cost of starting and and potential for a supposedly future Web
stopping application instances [43]. However, dubbed the 3D Web [55]. The global distribution
they fail to reduce server utilization in the worst- of such a vision is considerably more than just
case scenario; load balancing performance is not another evolutionary step in content visualization;
optimal. Furthermore, some apps, which we refer instead, future Internet infrastructure and
to as dependent applications, must connect; the protocols must address severe technical hurdles
network cost of these applications must also be [56]. Due to their reliance on centralized
considered [44]. solutions, existing commercial systems have
limited scalability. When considering global-scale
Distributed key-value systems (e.g., scenarios such as the 3D Web, distributed and
Memcached) are essential tools for caching decentralized approaches become essential.
popular content in memory, saving time and
money by avoiding complex and expensive Cloud computing results from the convergence of
database searches and file system visits [45]. It's some technologies that have changed the way
critical to balance the load across a cluster of businesses create IT infrastructure. Grid
cache servers if you want to make the most of computing, cluster computing, software-as-a-
your cache resources [46]. Current methods use service, utility computing, autonomic computing,
a proxy at each client to redirect requests across and other computing approaches are
the collection, but this necessitates client incorporated. It introduces an entirely new
modification and makes essential dynamic deployment mechanism for enterprise web apps
replication problematic. A centralized [57]. When opposed to employing an internal IT
representative can be utilized. However, this infrastructure, the cloud offers significant cost
hasn't always proven scalable [47]. savings. Cloud computing's "pay as you go"
concept is substantially less expensive for a firm
The Distributed Web Caching method allows for than internal IT's "pay for everything upfront"
quick retrieval of web pages, although server methodology [58]. Hardware Many cloud
delays still restrict it [48]. Due to highly infrastructure vendors' products rely on
congested servers, these systems have a low virtualization as an enabling technology. A
level of robustness. Servers are frequently physical server can be partitioned into many
disconnected, just as they are in real-time, virtual servers, each with its operating system
resulting in a service tradeoff. Robust Distributed and memory, CPU, and disk footprints, thanks to
Web Caching (RDWC) takes care of the virtualization [59].
robustness but not the frequent disconnections
[49]. The rapid rise of the Internet and the confluence
of computers and telecommunications have
Cluster administration is a time-consuming task. reshaped business regulations and the way
Allocating cluster resources by hand, in people communicate information. Users have
particular, can soon become unmanageable in a assured timeliness and reliability expectations for
dynamic environment like the Internet, where information service use in a varied and
processing requirements can vary rapidly [50]. developing Internet environment due to the rise
49
Ibrahim et al.; AJRCOS, 10(1): 47-62, 2021; Article no.AJRCOS.70184
of e-commerce [60]. These consumers' needs challenges, including data storage, retrieval, and
cannot be met by information systems based on manipulation [71]. The existing situation for
traditional client/server database technology [61]. managing the Indian Space Research
This necessitates creating an information service Organisation's (ISRO) ASTROSAT-CZTI satellite
system that can meet users' diverse service data involves an elementary and unsophisticated
expectations while also ensuring adaptability to technique, which results in significant data
deal with ever-changing events to give high retrieval delays and data portability concerns
assurance [62]. [72]. The complication, as mentioned
above's evaluation and research paved the path
Nowadays, the most common distributed for us to develop a new architectural
computing environments are Jini and.NET. The solution to handle this complex optimization
Jini architecture provides a framework for challenge [70].
defining, promoting, and discovering services in
a network, while.NET remoting allows developers The remainder of this article is structured as
to create Internet-based, distributed applications. follows: Section II Web server, section III Load
Both frameworks are capable of facilitating the balancing technology, section IV literature
implementation of distributed systems and are review, section V discussion and comparison,
incredibly comparable in many ways [63]. They and section VI is the conclusion.
are, however, incompatible. Jini clients cannot
request remote services from.NET systems, 2. WEB SERVER
and.NET clients are unable to locate Jini
services. Furthermore, no load-balancing A web server is a machine that can reply to client
methods are available in any of these situations requests. The response is regarding websites.
[64]. When software is installed and connected to the
Internet, a computer can become a web server
Routing is a crucial function that occurs at the [73, 74]. The web server is responsible for
network layer in any IP communication network. accepting and serving HTTP requests from web
Anycasting is a new mechanism for IP packet clients, usually in web pages containing static
delivery from a sending node to any of a group of content (text, images, etc.) and dynamic content
receiving nodes with the same IP addresses (scripts). Therefore, the leading role of the
introduced lately [65]. The sender will receive a webserver is to store and process files and
response from any one of the group's nodes. The provide clients with web page requests via a
introduction of IPv6 made it possible to locate communication protocol among clients and the
services using the anycasting approach, which server that uses the Hypertext Transfer Protocol
was not possible in IPv4. In most cases, packets (HTTP). The page submitted is a Hypertext
are routed using either a static or dynamic Markup Language (HTML) document. Many web
routing technique [66]. Single path routing (SPR), servers depended on the Apache web server,
multiple-path routing (MPR), and integrated which is the most commonly used web server, is
routing (SPR&MPR) approaches can be used to one of the most popular since Apache can be
route packets [67]. used. It is an HTTP web server and has access
control, PHP, and SSL support [75-78]. The most
In large-scale distributed virtual environments, popular web servers are Apache Web Server,
scalability and consistency are critical concerns Nginx, and NodeJS. The Apache web server is a
(DVEs). A VE is usually partitioned into several Unix-based, accessible, and open-source web
regions and handled by a series of dedicated server built by the Apache Software Foundation.
servers to make it scalable [68]. Because static Apache is lightweight, wholly featured, and more
arrangements can't handle dynamic loads, powerful than other web servers based on Unix.
dynamic methods are being researched and The architecture of Apache is thread-based,
tested extensively [69]. A multitude of solutions where the primary process (Multi-Processing
addressing either scalability or consistency may Modules-MPM) is named at startup and performs
be discovered in the literature, and systems child processes/threads (modules) to handle
requiring both demanded a new infrastructure requests simultaneously [79, 80]. Nginx,
[70]. developed by Igor Sysoev and then by NginX
Inc. in 2004, is a free, open-source, and high-
Satellites all across the world generate enormous performance web server. To benefit effectively
amounts of data regularly. The administration of from the hardware resources available, NginX
such massive amounts of data presents some often uses multi-processes.
50
Ibrahim et al.; AJRCOS, 10(1): 47-62, 2021; Article no.AJRCOS.70184
Together with the cache loader, a master provide high performance by preventing each
process is called at startup to load disk-based server from being overloaded inside the cluster.
cache in memory and cache manager to monitor In addition, the load balancer is a means of
memory utilization. This design seeks to reduce software or hardware designed to balance traffic
the impact of context switching in the architecture across various servers and brilliantly share client
of multi-processes. NodeJS is a JavaScript load across several nodes to one real IP [87, 88].
single-threaded server-side environment created Load balancing extends the bandwidth of
in 2009 by Ryan Dahl. With a scalable network devices, and servers improve throughput
architecture based on events, NodeJS can and improve network data processing capabilities
achieve a high concurrency level. However, for [89]. To ensure high efficiency and high-
non-blocking I/O operations, NodeJS also uses performance computing of the system, some
several threads. Many of NodeJS's main reasonable resource allocation is based on the
modules are written in Java script and run on the load pressure [90, 91]. In addition, server load
Google V8 Java script Engine. Thus, NodeJS's balancing provides applications, cloud services,
entire code path is asynchronous and non- and websites with scalability and high availability
blocking [76, 79, 81, 82]. by monitoring server health, distributing loads
equally among servers, and maintaining session
3. LOAD BALANCING TECHNOLOGY persistence if one or more servers are
overloaded or unresponsive. Several load
Web traffic has generally affected the balancing algorithms are used to balance the
performance of any web server, and it makes a workload, such as round robin, Least
slow response of web server due to get Connections, and Weighted Least Connections
overloaded. Each web server faces the problem [92, 93]. The Load balance can be divided into
of overloading and needs an optimal solution. So software and hardware balancing methods. The
to overcome this, using load balancing first technology (Software) is done by installing
techniques [83, 84]. It acts as a transmitter on one or many nodes (servers) the
between the requests of the client and the web corresponding additional software and
servers by distributing the incoming requests to performing the complementary software load
all the nodes within the cluster to improve the balancing technique. The benefits of the first
use of speed and resources and ensure cost- technology are easy to configure, flexible use, no
effective and scalable [85, 86]. It is also used to external device needed, and low cost.
51
Ibrahim et al.; AJRCOS, 10(1): 47-62, 2021; Article no.AJRCOS.70184
The drawback of the software load balancer is improve the load balancing capability of the
the program itself can consume some machine webserver. First, for all the performance
resources. The problem of the resource usage of indicators of each server in the cluster, the
the program itself will also become a bottleneck quantitative values are specified. The initial
of the load capacity of the device when the virtual node set is measured according to the
number of requests rises to a certain degree. In quantization value, optimizing the unequal load
general, loading in the form of hardware installs distribution caused by differences in the server's
the load balancer directly to the server and the performance. Then the load is refined, and the
output connected to the external network. It is virtual node is dynamically calculated based on
independent of the OS and uses special the measured server's performance value and
installation facilities, which dramatically improves load value so that the cluster load is more
enhancement. The drawback is that the cost of balanced. The result shows that the proposed
hardware upgrades is costly [19, 94]. The method minimizes the average response time
classification of load balancing illustrates in and maximizes throughput.
Fig. 1.
In 2018, Putu Adhi Suwandika and et al. [99]
4. LITERATURE REVIEW proposed Least Connection and IP Hash load
balancing algorithm to be applied on web servers
In 2017, Pan Zhu and Jiangxing Zhang [95], over SDN networks. Both algorithms will be
Improved Load Balancing approach for Web evaluated with some basic parameters for a web
Server Cluster Enhanced Weighted Minimum server. Response Time, Throughput, and
Join. Firstly, the algorithm considers the usage of Resource Utilization are these parameters. The
the server, memory use, and mutilation of parameter test results show that the IP Hash
network bandwidth. Other variables are algorithm provides 17% more optimal
combined with the web server's properties to performance response time, 10% more optimal
dynamically change the weight's size according throughput, and 8% more effective memory use
to the server's real-time load. Secondly, assigns than the Least Connection algorithm.
the new connection request according to the
weighted minimum connection number algorithm. In 2018, Guoqing Liu and et al. [100], A new load
The algorithm reduces server latency, decreases balancing algorithm was proposed that considers
HTTP response time, and can effectively the nodes' load and performance information and
increase load balancing performance using the type of request to solve the problem of
OPNET simulation tools. unequal load distribution in cluster servers. The
discussed approach achieves a relatively
In 2017, Peirui JIANG and et al. [96], a novel optimum load balance compared to the
load balancing algorithm was suggested that traditional Round-Robin algorithm.
considers the proximity between client and
server. The Client Proximity Based Load In 2018, Gurasis Singh and Kamalpreet Kaur
Balancing Algorithm (PBLB) is compared with the [101] enhanced the weighted least connection
Modified Random Allocation (MRA) method and algorithm, which manages the load among
the Round Robin (RR) scheme. During the different servers by preventing requests to the
simulation, the PBLB algorithm has a higher new server. The proposed algorithm excludes
performance as compared to MRA and RR and disables a new server from its scheduling list
algorithms. when HTTP requests are continuously allocated
to only a new server above the maximum
In 2017, Sanjaya Kumar Panda and et al. [97] continuous assigned number (C). Finally, after C-
proposed (SIS) algorithm to distribute and 1 allocation round times, the new server is
balance the load among servers, using a three activated and included in the server scheduling
steps process. The SIS algorithm is compared list. The suggested algorithm balances the load
with the client/server approach, and the between real servers by preventing overloads on
performance is evaluated according to the the latest real server.
number of load comparisons and the load factor.
The obtained result shows the performance of In 2018, Iqsyahiro Kresna A and Yusep
the algorithm proposed. Rosmansyah [102] implemented a web server
cluster and load balancer as a high availability
In 2018, Jie Li and et al. [98] A DCH algorithm framework, which is expected to overload the
based on consistency hash were suggested to server problem. The results of the load balancer
52
Ibrahim et al.; AJRCOS, 10(1): 47-62, 2021; Article no.AJRCOS.70184
test showed that the LC algorithm was request to the lightest pack on the webserver.
outperforming the RR algorithm. In High Obtained results illustrated that the algorithm
Availability or failover experimenting, the average increases software and hardware resource
downtime is 6,587 seconds. usage and balances the load among servers.
In 2018, Zepeng Wen and et al. [103] presented In 2018, Ruoyu Li and et al. [107], Analyses the
a dynamic feedback load balancing algorithm container-based WEB application deployment
based on Nginx. The proposed algorithm is architecture, measures the host load parameter
realized based on queuing theory and through combined with the performance of the host
the plug-in mechanism of Nginx by developing an indicates and operating container status,
algorithm. A test environment was then including Processor use, memory use, network
designed. Loadrunner software is used to use, and the amount of unused memory
perform a comparison test on the efficiency of assigned to the containers, and suggests a
the Weighted Round-Robin Scheduling, IP Hash Dynamic Weighted Least-Connection Algo
algorithm, and the dynamic feedback algorithm (DWLC). The results show that the DWLC
built-in by Nginx. The results showed that the algorithm is 52.6 % and 46.4 % faster than the
load balance could be better with the dynamic ordinary Round-Robin and Least-Connection
algorithm used in this work. algorithm for the WEB application response.
In 2018, Deepti Sharma [104] presented a load In 2018, Fatma Mbarek and Volodymyr Mosorov
balancing approach that allows HTTP requests to [108], load balancing was dedicated to handling
be distributed dynamically among all servers in the web cluster for distributed and parallel
the cluster depend on the response time. In this systems. Using several scheduling algorithms, it
work, the load balancer performs on a lightweight distributes the load across web servers. This
server that is easily integrated with java. study discusses a detailed comparison of several
Requests come from two sources, either from the heterogeneous web cluster HAProxy algorithms
browser or from another tool called SoapUII. A and the metaheuristic methods inspired by the
track of HTTP requests and responses on all behavior of the insect colonies, such as the Ant
servers is kept in the hash data structure. Due to Colony Optimization algorithm and the Honey
the response time and pending workloads for Bee algorithm. Solving combinatorial optimization
that server, a new request is distributed on web problems is the fundamental concept of meta-
servers. Different experimental results show the heuristic methods.
efficiency of the proposed algorithm.
In 2018, Minato Omori and Hiroaki Nishi [109]
In 2018, Hsien-Yi Liu and et al. [105] suggested proposed a load balancing algorithm that
an Open Flow-based Least-Loading load- distributes the request based on processing time
balancing technique to make full use of hardware estimates, which prevents mismatches between
resources. The proposed algorithm allocates server characteristics and the content of the
requests to the server and every node in the request. Based on the requested content, the
server pool's real-time status. Furthermore, processing time for submissions is estimated by
minimizing the overhead of choosing the request online machine learning. A technique to cover
managing server and essentially increasing the the latency of machine learning is proposed and
entire system throughput, the least loading partially carried out. They developed a model of
strategy prevents the controller from being a multiple database servers to test the algorithm
system bottleneck. Experimental findings show and performed an experiment using actual log
that the Lest Loading strategy is 17.2% and data for database requests. The simulation
21.4% better than RR and WRS algorithms in the results indicate that, compared to round-robin,
Floodlight controller. the proposed algorithm decreased the average
processing time for requests by 94.5 % and by
In 2018, Anji Yu and Shimin Yang [106], a new 28.3 % to the least connections.
approach that distributes a load on the server’s
algorithm, is suggested to minimize the execution In 2018, Mochamad Rexa Mei Bella a dit al [110],
time. The algorithm's strategy is to divide the A method was proposed to track each host
degree of effect on the request client's server machine's memory utilization and distribute web
and assign the corresponding weight. The traffic based on the memory utilization of each
performance and load of the server are host machine. Promising is the outcome of this
combined at the same time to transfer the experiment. Each of the worker nodes receives
53
Ibrahim et al.; AJRCOS, 10(1): 47-62, 2021; Article no.AJRCOS.70184
Table 1. Summary related to the previous study (webserver with load balancing technique)
54
Ibrahim et al.; AJRCOS, 10(1): 47-62, 2021; Article no.AJRCOS.70184
55
Ibrahim et al.; AJRCOS, 10(1): 47-62, 2021; Article no.AJRCOS.70184
fair load distribution. This technique can minimize The summary of the work performed by
the probability of a single point of failure in the researchers is explained in the above section
web server cluster. after applying load balancing algorithms to
improve web server performance illustrated in
In 2019, Kadiyala Ramana and M. Ponnavaikko Table 1.
[111], An approximate web server queuing
technique was introduced for web server clusters 5. ASSESSMENT AND DISCUSSION
and an analytical model for web server load
calculation. To achieve better efficiency, the This research aims to overview the dynamic load
requests are categorized according to the service balance techniques to improve web server
time and track the number of outstanding performance. The dynamic load balancing
requests on each webserver. The estimated load algorithm adopted will allocate the vast data
of each web server is used for load balancing. request content equally to each server and
The research findings demonstrate the efficacy enhance the server's concurrent processing
of the suggested approach by enhancing the capacity to shorten the average server response
server cluster's mean response time, throughput, time and increase the best degree of adaptation.
and drop rate. According to the previous studies related to
improving the performance of web servers by
In 2020, E Qin and et al. [112]. Improved Nginx- using a dynamic load balance technique
based dynamic weighted round-robin algorithm, mentioned in the last section, many algorithms
which collects real-time server load information are used to enhance the work of web servers,
and dynamically adjusted weight. The proposed and most of them achieve good performance.
algorithm is compared to weighted round-robin
approaches in terms of throughput and response The proposed algorithm (pending job algorithm)
time. The obtained result demonstrates that the in [104] has been validated using a different
proposed method performs better than the number of requests ranging from 100 to 10000.
weighted round-robin algorithm. The algorithm sets the server priority with the
lowest response time and the lowest pending
In 2020, WEI Chunlei and et al. [113] proposed request. So each server will handle a different
an approach based on the socket buffer's number of requests. This algorithm is analyzed
feedback mechanism, which takes the server as based on two parameters. First, the proposed
the load indicator, collects the load indicator via algorithm manages transactions per second,
the load agent, and calculates the discrete server handling more transactions that achieve better
socket buffer coefficient. The dispatch controller efficiency. The second parameter is the total
determines the weight of each server, and the processing time, and the proposed algorithm
higher the weight, the lighter the server load is. takes less time to complete all jobs than other
The tests illustrate that the CMTS feedback on algorithms. The author in [99] depended on using
the server load status is more realistic than the two algorithms, IP hash and the least connection
LC method. The average cluster response time is algorithms. The two algorithms are evaluated
reduced as the number of cluster node using three different parameters. The first
connections and message sending rates to parameter is response time, and the response
maximize. time testing is done by using wget tools when 20
clients send a request to the server. With the IP
In 2020, Mei-Ling Chian and et al. [114], hash technique, the average response time
Dynamic weighted random selection was indicates the desired result. The second
proposed as a novel load balancing algorithm parameter is throughput. Perf tools tease the
(DWRS). DWRS considers the real-time server throughput of the IP hash algorithm. When
loads while delivering requests to servers. implementing the load balancing approach with a
Underutilized servers are given larger weights, predetermined algorithm, throughput testing is
increasing their chances of being chosen to performed to evaluate the bandwidth efficiency
process requests. Modify the packet handling on the webserver. The IP Hash algorithm shows
flow in the Floodlight controller to improve stable performance on all servers. The third
system performance. When selecting the target parameter is resource utilization. Htop tools do
server, a multi-threaded technique is employed this parameter testing to retrieve current memory
to effectively exploit the parallel processing usage. Testing is done with ten clients up to 100
power of numerous cores, which speeds up the clients. According to the obtained result, the IP
processing of packet-in messages. hash algorithm is more suitable than the
56
Ibrahim et al.; AJRCOS, 10(1): 47-62, 2021; Article no.AJRCOS.70184
57
Ibrahim et al.; AJRCOS, 10(1): 47-62, 2021; Article no.AJRCOS.70184
study of moving from cloud computing to 22. Zebari SR, Yaseen NO. Effects of Parallel
fog computing. Qubahan Academic processing implementation on balanced
Journal. 2021;1:60-70. load-division depending on distributed
12. Abdulrahman LM, Zeebaree SR, Kak SF, memory systems. J. Univ. Anbar Pure Sci.
Sadeeq MA, Adel AZ, Salim BW. et al. A 2011;5:50-56.
state of art for smart gateways issues and 23. Shukla P, Kumar A. CLUE based load
modification. Asian Journal of Research in balancing in replicated web server. in 2018
Computer Science. 2021;1-13. 8th International Conference on
13. Ramana K, Ponnavaikko M. A Multi-Class Communication Systems and Network
Load Balancing Algorithm (MCLB) for Technologies (CSNT). 2018;104-107.
heterogeneous web cluster. Stud. Informat. 24. Aziz ZAA, Ameen SYA. Air pollution
Control. 2018;27:443-452. monitoring using wireless sensor
14. Hassan RJ, Zeebaree SR, Ameen SY, Kak Networks. Journal of Information
SF, Sadeeq MA, Ageed ZS, et al. State of Technology and Informatics. 2021;1:20-25.
art survey for iot effects on smart city 25. Yazdeen AA, Zeebaree SR, Sadeeq MM,
technology: challenges, opportunities, and Kak SF, Ahmed OM, Zebari RR. FPGA
solutions. Asian Journal of Research in implementations for data encryption and
Computer Science. 2021;32-48. decryption via concurrent and parallel
15. Yusuf R, Nuha RSPU. Comparative computation: A review. Qubahan
analysis of HAProxy& Nginx in round robin Academic Journal. 2021;1:8-16.
algorithm to deal with multiple web 26. Abdulmohsin HA. A load balancing
request; 2018. scheme for a server cluster using history
16. Yasin HM, Zeebaree SR, Sadeeq MA, results. Iraqi Journal of Science.
Ameen SY, Ibrahim IM, Zebari RR, et al. 2016;57:2121-2130.
IoT and ICT based smart water 27. Amanuel SVA, Ameen SYA. Device-to-
management, monitoring and controlling device communication for 5G security: A
system: A review. Asian Journal of review. Journal of Information Technology
Research in Computer Science. 2021;42- and Informatics. 2021;1:26-31.
56. 28. Elzanati WM, Ameen SY. Cost effective
17. Malallah H, Zeebaree SR, Zebari RR, air-conditioning for bahrain domestic
Sadeeq MA, Ageed ZS, Ibrahim IM, et al. applications. in 2013 7th IEEE GCC
A comprehensive study of kernel (Issues Conference and Exhibition (GCC).
and Concepts) in different operating 2013;535-540.
systems. Asian Journal of Research in 29. Ameen SY, Al-Badrany MR. Optimal image
Computer Science. 2021;16-31. steganography content destruction
18. Ageed ZS, Zeebaree SR, Sadeeq MA, techniques. in International Conference on
Abdulrazzaq MB, Salim BW, Salih AA, et Systems, Control, Signal Processing and
al. A state of art survey for intelligent Informatics. 2013;453-457.
energy monitoring systems. Asian Journal 30. Fawzi LM, Alqarawi SM, Ameen SY,
of Research in Computer Science. Dawood SA. Two levels alert verification
2021;46-61. technique for Smart Oil Pipeline
19. Li B, Shang J, Dong M, He Y. Research Surveillance System (SOPSS),.
and application of server cluster load International Journal of Computing and
balancing technology. in 2020 IEEE 4th Digital Systems. 2019;8:115-124.
Information Technology, Networking, 31. Ameen SY, Ahmed IM. Design and
Electronic and Automation Control implementation of e-laboratory for
Conference (ITNEC). 2020;2622-2625. information security training. in 2013
20. Gebrehiwot ME, Aalto S, Lassila P. Energy Fourth International Conference on e-
efficient load balancing in web server Learning. Best Practices in Management,
clusters," in 2017 29th International Design and Development of e-Courses:
Teletraffic Congress (ITC 29). 2017;13-18. Standards of Excellence and Creativity.
21. Abdullah SMSA, Ameen SYA, Sadeeq MA, 2013;310-317.
Zeebaree S. Multimodal emotion 32. Ameen SY, Al-Jammas MH, Alenezi AS.
recognition using deep learning. Journal of FPGA implementation of modified
Applied Science and Technology Trends. architecture for adaptive Viterbi decoder. in
2021;2:52-58. 2011 Saudi International Electronics,
58
Ibrahim et al.; AJRCOS, 10(1): 47-62, 2021; Article no.AJRCOS.70184
Communications and Photonics 44. Fawzi LM, Ameen SY, Alqaraawi SM,
Conference (SIECPC). 2011;1-9. Dawwd SA. Embedded real-time video
33. Al-Sultan MR, Ameen SY, Abduallah WM. surveillance system based on multi-sensor
Real time implementation of stegofirewall and visual tracking. Appl. Math. Infor. Sci.
system. International Journal of Computing 2018;12:345-359.
and Digital Systems. 2019;8:498-504. 45. Rashid ZN, Zeebaree SR, Shengul A.
34. Ameen SY, Mahdi AH. AES cryptosystem Design and analysis of proposed remote
development using neural networks. controlling distributed parallel computing
International Journal of Computer and system over the cloud. in 2019
Electrical Engineering. 2011;3:309. International Conference on Advanced
35. Ameen SY, Ibrahimi IA. MANET routing Science and Engineering (ICOASE).
protocols performance evaluation with TCP 2019;118-123.
Taho, Reno and new-Reno. International 46. Rashid ZN, Zeebaree SR, Sengur A. Novel
Journal of u-and e-Service, Science and remote parallel processing code-breaker
Technology. 2011;4:37-49. system via cloud computing."
36. Ameen SY, Saud LJ. Computing nodes 47. Ali ZA, Ameen SY. Detection and
and links appearances on geodesics in prevention cyber-attacks for smart
network topologies using graph theory. in buildings via private cloud environment.
Full Papers Conference Proceeding. International Journal of Computing and
2011;235. Network Technology. 2018;6:27-33.
37. Al Janaby AO, Al-Omary A, Ameen SY, Al- 48. Rashid ZN, Zebari SR, Sharif KH, Jacksi
Rizzo HM. Tracking high-speed users K. Distributed cloud computing and
using SNR-CQI mapping in LTE-A distributed parallel computing: A review. in
networks. in 2018 International Conference 2018 International Conference on
on Innovation and Intelligence for Advanced Science and Engineering
Informatics, Computing, and Technologies (ICOASE). 2018;167-172.
(3ICT). 2018;1-7. 49. Farhan FY, Ameen SY. Improved hybrid
38. Taher KI, Saeed RH, Ibrahim RK, Rashid variable and fixed step size least mean
ZN, Haji LM, Omar N, et al. Efficiency of square adaptive filter algorithm with
semantic web implementation on cloud application to time varying system
computing: A review. Qubahan Academic identification. in 2015 10th System of
Journal. 2021;1:1-9. Systems Engineering Conference (SoSE).
39. Othman A, Ameen SY, Al-Rizzo H. 2015;94-98.
Dynamic switching of scheduling algorithm 50. Rashid ZN, Sharif KH, Zeebaree S.
for. International Journal of Computing and Client/servers clustering effects on CPU
Network Technology. 2018;6. execution-time, CPU usage and CPU Idle
40. Sadeeq MM, Abdulkareem NM, Zeebaree depending on activities of parallel-
SR, Ahmed DM, Sami AS, Zebari RR. IoT processing-technique operations.
and Cloud computing issues, challenges Int. J. Sci. Technol. Res. 2018;7:106-
and opportunities: A review. Qubahan 111.
Academic Journal. 2021;1: 1-7. 51. Fawzi LM, Ameen SY, Dawwd SA,
41. Ameen SY, Ali ALSH. A comparative study Alqaraawi SM. Comparative study of ad-
for new aspects to quantum key hoc routing protocol for oil and gas
distribution. Journal of Engineering and pipelines surveillance systems.
Sustainable Development. 2018;11:45-57. International Journal of Computing and
42. Kareem FQ, Zeebaree SR, Dino HI, Network Technology. 2016;4.
Sadeeq MA, Rashid ZN, Hasan DA, et al. 52. Jijo BT, Zeebaree SR, Zebari RR, Sadeeq
A survey of optical fiber communications: MA, Sallow AB, Mohsin S, et al. A
challenges and processing time influences. comprehensive survey of 5G mm-wave
Asian Journal of Research in Computer technology design challenges. Asian
Science. 2021;48-58. Journal of Research in Computer Science.
43. Omer MA, Zeebaree SR, Sadeeq MA, 2021;1-20.
Salim BW, x Mohsin S, Rashid ZN, et al. 53. Sadeeq MA, Zeebaree S. Energy
Efficiency of malware detection in android management for internet of things via
system: A survey. Asian Journal of distributed systems. Journal of Applied
Research in Computer Science. 2021;59- Science and Technology Trends.
69. 2021;2:59-71.
59
Ibrahim et al.; AJRCOS, 10(1): 47-62, 2021; Article no.AJRCOS.70184
54. Othman A, Ameen SY, Al-Rizzo H. A new gamification examination fields. Technol.
channel quality indicator mapping scheme Rep. Kansai Univ. 2020;62:2319-2328.
for high mobility applications in LTE 64. Ameen SY, Nourildean SW. Firewall and
systems. Journal of Modeling and VPN investigation on cloud computing
Simulation of Antennas and Propagation. performance. International Journal of
2015;1:38-43. Computer Science and Engineering
55. Maulud DH, Zeebaree SR, Jacksi K, Survey. 2014;5:15.
Sadeeq MAM, Sharif KH. State of art for 65. Sadeeq M, Abdulla AI, Abdulraheem AS,
semantic analysis of natural language Ageed ZS. Impact of electronic commerce
processing. Qubahan Academic Journal. on enterprise business. Technol. Rep.
2021;1:21-28. Kansai Univ. 2020;62:2365-2378.
56. Shukur H, Zeebaree SR, Ahmed AJ, 66. Alzakholi O, Shukur H, Zebari R, Abas S,
Zebari RR, Ahmed O, Tahir BSA, et al. A Sadeeq M. Comparison among cloud
state of art survey for concurrent technologies and cloud performance.
computation and clustering of parallel Journal of Applied Science and
computing for distributed systems. Journal Technology Trends. 2020;1:40-47.
of Applied Science and Technology 67. Al-Khayat ON, Ameen SY, Abdallah MN.
Trends. 2020;1:148-154. WSNs power consumption reduction using
57. Jacksi K, Ibrahim RK, Zeebaree SR, clustering and multiple access techniques.
Zebari RR, Sadeeq MA. Clustering International Journal of Computer
documents based on semantic similarity Applications. 2014;87.
using HAC and K-Mean algorithms. in 68. Ageed Z, Mahmood MR, Sadeeq M,
2020 International Conference on Abdulrazzaq MB, Dino H. Cloud computing
Advanced Science and Engineering resources impacts on heavy-load parallel
(ICOASE). 2020;205-210. processing approaches. IOSR Journal of
58. Sadeeq MA, Abdulazeez AM. Neural Computer Engineering (IOSR-JCE).
networks architectures design, and 2020;22:30-41.
applications: A review. in 2020 69. Sallow A, Zeebaree S, Zebari R, Mahmood
International Conference on Advanced M, Abdulrazzaq M, Sadeeq M. Vaccine
Science and Engineering (ICOASE). tracker," SMS reminder system: Design
2020;199-204. and implementation; 2020.
59. Othman A, Othman SY, Al-Omary A, Al- 70. Ameen SY, Yousif MK. Decode and
Rizzo H. Comparative performance of forward cooperative protocol enhancement
subcarrier schedulers in uplink LTE-A using interference cancellation. Int. J.
under high users' mobility. International Elect., Comput., Electron. Commun. Eng.
Journal of Computing and Digital Systems. 2014;8:273-277.
2015;4. 71. Sadeeq MA, Zeebaree SR, Qashi R,
60. Ageed ZS, Ibrahim RK, Sadeeq MA. Ahmed SH, Jacksi K. Internet of things
Unified ontology implementation of cloud security: A survey. in 2018 International
computing for distributed systems. Current Conference on Advanced Science and
Journal of Applied Science and Engineering (ICOASE). 2018;162-166.
Technology. 2020;82-97. 72. Abdulazeez AM, Zeebaree SR, Sadeeq
61. Sallow AB, Sadeeq M, Zebari RR, MA. Design and implementation of
Abdulrazzaq MB, Mahmood MR, Shukur electronic student affairs system.
HM, et al. An investigation for mobile Academic Journal of Nawroz University.
malware behavioral and detection 2018;7:66-73.
techniques based on android platform. 73. Prabu U, Malarvizhi N, Amudhavel J,
IOSR Journal of Computer Engineering Sriram R, Ravisasthiri P. Load balancing
(IOSR-JCE). 22:14-20. policies of web servers: research analysis,
62. Othman A, Ameen SY, Al-Rizzo H. An classification and perspectives. EAI
energy-efficient MIMO-based 4G LTE-A Endorsed Transactions on Scalable
adaptive modulation and coding scheme Information Systems. 2019;6.
for high mobility scenarios. International 74. Abdullah DM, Ameen SY. Enhanced
Journal of Computing and Network Mobile Broadband (EMBB): A review.
Technology. 2015;3. Journal of Information Technology and
63. Sulaiman MA, Sadeeq M, Abdulraheem Informatics. 2021;1:13-19.
AS, Abdulla AI. Analyzation Study for
60
Ibrahim et al.; AJRCOS, 10(1): 47-62, 2021; Article no.AJRCOS.70184
61
Ibrahim et al.; AJRCOS, 10(1): 47-62, 2021; Article no.AJRCOS.70184
Peer-review history:
The peer review history for this paper can be accessed here:
http://www.sdiarticle4.com/review-history/70184
62