MODEL AND ALGORITHM FOR INCREASING THE EFFICIENCY OF REMOTE SERVICE SYSTEMS SERVER

Advanced Computing: An International Journal (ACIJ), Vol.15, No.1/2/3/4, July 2024
DOI:10.5121/acij.2024.15402 11
MODEL AND ALGORITHM FOR INCREASING THE
EFFICIENCY OF REMOTE
SERVICE SYSTEMS SERVER
Vakhid Zakirov and Eldor Abdullaev
Department of Radioelectronic devices and systems, Tashkent state transport university,
Tashkent, Uzbekistan
ABSTRACT
This article examines the aspects that influence productivity in distant service systems during times of
request congestion. At the same time, the issue of enhancing system server efficiency without the use of
additional hardware or software has been investigated. It has been found that service system efficiency
declines during traffic congestion because requests are processed in multiple stages, each with a different
service order. Hence, to enhance system efficiency during traffic congestion, a proposal is made to
prioritize users whose requests have been successfully served at least once. A physical model of the service
process was constructed to assess the efficacy of the proposed approach.The research conducted using this
model contributed to an average increase in the efficiency of service systems by 6%.
KEYWORDS
server performance indicator, priority service, request traffic, service algorithm, mass service model,
service time
1. INTRODUCTION
Our dependence on web applications is steadily deepening, permeating various facets of our
lives. These platforms serve as conduits for a myriad of services across different social strata,
fueling a surge in their usage. Consequently, this heightened utilization brings forth a host of
systemic challenges [7], [12], [29]. Specifically, as user numbers and requests increase, the
system grapples with escalating request traffic, resulting in diminished server performance
metrics. Hence, optimizing server efficiency amidst traffic congestion emerges as a paramount
concern [2], [5], [13], [30].
The significance of the situation has resulted in the carrying out of numerous studies in this area.
Such studies were conducted in the following directions: implementation of various hardware and
code optimizations in the service of requests [4], [20], [21], [27], server service times for request
reduction [18], load balancing [22], [25], [26], service time prediction [28], web server
performance evaluation [23], [24].
The focus of research into various optimizations in the service of requests is to improve the user
interface, reduce server load, and increase system interaction through the use of FrontEnd,
Backend, and database technologies. These optimization methods have been built in order to
reduce latency and increase system efficiency.

12
The research aimed to reduce server request service times by utilizing various online optimization
methods, such as Newton's Method, Fuzzy Control, Heuristic Approaches, Feedback
Mechanisms, and Qualitative Comparisons [4], [20], [21], [27]. These optimizations were applied
to the hardware, software backend, and frontend components.Additionally, the research examined
the impact of technical device configurations, including browser caching, DNS caching, HTTP
keep-alive timeout, and gzip compression, on their effectiveness and the dynamic changes in
request volume.
The research also explored load balancing techniques to improve system efficiency. The goal was
to equally distribute the incoming requests and their associated load among the service servers
[22], [25], [26].
The research also involved predicting service times and evaluating web server performance. Web
monitoring tools were used to analyze how web server performance depends on factors such as
the distance between the server and the user, as well as the size of the web page [23], [24], [28].
Unlike earlier researches, this study focuses on developing algorithms and models to modify the
order of service requests when the system becomes congested. The purpose is to handle
circumstances where traffic jams develop in the service system, which were not addressed by the
previous activities stated.
The research work consists of five sections, conclusions, and a reference list. The analysis of the
reasons for the problem is presented in the second section, the methods used in the research in the
third section, the experimental setup section is located in the fourth section, and the results and
discussions of the research in the fifth section.
2. PROBLEM STATEMENT
The problem of request congestion in the system usually occurs during times of high intensity of
requests entering the system. Also, the origin of traffic directly depends on the service algorithm
of the system. Currently, requests are served using a web server in the order shown in Figure 1.
Control process
Server Thread
Server Thread
Server Thread
Server Thread
Server Thread
Server Process
Server Process
Listen
Main Thread
Listener Thread
TCP/IP
listen
queue
Connection
form clients
HTTP server
Server Process
ThreadsPerChild
ListenBackLog
OS
Database
Startup and
monitoring
Figure 1. The way the web server serves requests
Users typically request multiple times from the system in order to receive the one crucial piece of
information. In this instance, the system's response time to the user's first and subsequent requests
will differ. In this instance, the user's initial request is authenticated by the system.Authentication

13
necessitates a sequence of checks between the user and the system, including seeking up the
identification information in the system database, verifying its validity, and then allocating a
service device to the user, which takes more time to service requests. This technique may also be
associated with receiving data from external devices. This procedure is not repeated for further
user requests for this information, unless the request interval exceeds the system's user-assigned
session time. As a result, the time taken to service these two types of requests is different. As a
result, during the identification process, the system provides no helpful service to the user and
consumes system resources inefficiently. This creates various delays in allowing new users into
the system and serving other requests into the system when there is traffic, resulting in a drop in
system efficiency.
Requests that have successfully passed the identification process will continue to be served if the
service devices are empty.
If all service devices are busy, they will be queued. At the same time, since the number of system
service devices and request queues has a certain threshold value, it causes them to be quickly
filled with requests during times of high traffic.When the service devices are empty, the system
can serve the requests in the waiting areas in different order. One of these FIFO (First in First
out), LIFO (Last in First out) or SIRO (Service in random order) methods can be used [16]. In
our view, requests in waiting areas are served on a FIFO basis. In the process, the system does
not care whether the user has passed the identification or not and accepts the request for
service.This causes the user's requests to remain in the queue more often or to be excluded from
the service. Because of this, the system's efficiency declines.
As a consequence, developing appropriate approaches, models, and algorithms for solving these
types of challenges is one of today's most critical issues.. At the same time, troubleshooting
without using additional hardware or software tools does not increase the system server load and
does not slow down the service process. Therefore, it is important to research service algorithms,
models, and optimization processes to achieve positive results.
3. METHODS
Taking into account the difference in service time for the initial and subsequent requests
mentioned above gives some degree of results when solving problems of service requests during
high-traffic times.For this, it is necessary to divide the requests into two types. The first type
includes requests that need to go through the identification process. The latter includes requests
that have previously successfully passed identification to obtain a single piece of information
from this process. It allows to divide the requests into two types, to serve them in different order
during times of high traffic. This reduces the time it takes to service a request and thus increases
the number of fully serviced requests.
We can use the following model of mass service to divide incoming requests into two and serve
them in different order [14], [15], [16], [17](Figure 2). This model is a service model based on
the priority of requests during traffic congestion.

14
1
2
3
j
...
1
2
3
n-j
...
1 2 3 m
...
r . . . 4 3 2 1
Queue of
non-
privileged
requests
Queue of
privileged
requests
Service devices
H
1-H
Sign out
b
d
c
f
Log the user out of
the system
a
K
P
𝐼𝐼 = 𝛼 ∗ 𝑗
𝐼 = 𝛼 ∗ (𝑛 − 𝑖 − 𝑗)
Q1
Q2
Figure 2.A service model based on the priority of requests during traffic congestion
In this process, two groups of K and P - unprivileged and privileged requests are coming to the
system. The first K group of users form the initial request stream. They are equal to
)
( j
i
n
I 



 the total rate, and the  incoming requests from the n source, each of which
has a rate, create flows.The second P group is formed by the flow of requests successfully served
at least once by the system II
 =𝛼 ∗ 𝑗.
A common request stream from these two sources K and P is served in standby mode
II
I 

 

 by a system of service devices m (e.g., a server).The total number of request
queues is r .
In this case, service of requests is carried out in two stages - preliminary and main. At the initial
stage, group K requests go through the identification process. In the process of identification, an
HTTP request, information exchange with the DNS server, etc. are carried out. At the second
stage, the necessary information exchange is carried out and the requests are fully serviced. If the
user takes the necessary information for itself and wants to complete his work, then he will log
out with a probability of 1-H.
A request received through the first K group goes through the request identification process when
the service devices are free. In the second step, the service device (server) prepares the desired
result for the request and sends it to the user. If the process is completed successfully, the source
is assigned a certain symbol - the "priority" symbol, and the source is included in the group P
with probability H. Such sources form the second group of sources. When traffic is not observed
in the system, all K and P group requests are served in the same order. In times of traffic
congestion, requests with a priority mark are served with a certain privilege.This privilege
consists of the following. If there is an empty service device at the moment of time when the
request falls, then such a request is served immediately without an identification process, and the
service is carried out directly through channel b in Figure 2.If there is no free service device
available at the moment of the request, privileged requests are transferred to the request queue
through the d channel and are served in the FIFO method [2]. Requests in non-privileged queues
are serviced after privileged requests have been serviced.
If all the service devices are busy at the moment of the request from the first K group source, the
request is transferred to the standby devices through channel c. If the queues are busy, the request

15
will be rejected.If all service devices are busy when a request is received from the second source
P, the request is transferred to the queue devices. If all the queues are busy, the request from the
first K source in the queue is rejected through channel f, and the second source request is placed
in the queue.
Also, this process is terminated when the service devices of the system are free, and the service of
requests from the source K is continued. It should also be noted that there is a number of requests
that can be received simultaneously from groups K and P, and these values are equal to n in
total.Here, n is equal to the sum of service devices - m and the number of queues - r. In this case,
when there are no users in the system at all, n values belong to K groups. However, as soon as
there are users who successfully used the system at least once, n values are given to P group. The
number of users in group P is j.And when it is equal to the value of j, incoming requests in the
system are accepted only through group P, and all users are considered users with priority. The
algorithmic view of this model is as follows (Figure 3).
Begin
K and P
channel
requests
(n-r)<m
Accepting the
request for service
through channels a
and b
Yes
m-1
Break the
request into
parts
Send a request
to the software
Accepting a request
in the software
Checking that the
user has been
authenticated
Formation of a
survey result
Yes
Transferring the user
to the identification
process
No
Accepting login
and password
from the user
Checking the
validity of the
login password
Yes
Send a message to the user
about a login password error
No
Attach a priority
symbol to a user
Grouping P with
probability H
Send the result
to the user
(1-H)
The end
n>r+m
No
Put the request on hold
through channels c and
d
Yes
r-1
Remove request
from service
No
r>n-m
Acceptance of the first
request for service in the
FIFO method
through channel Q1
Yes
Does the request
have priority?
Accepting the first
request in the FIFO
method
No
Acceptance of
the request for
service through
the Q2 channel
Yes
Denial of
service through
channel f
No
Figure 3. Algorithm of priority service during traffic congestion

16
4. EXPERIMENTAL SETUP
To solve the models of this type and to determine the quality indicators of the system (waiting
probability, waiting time, etc.), it is possible to use two methods - strictly mathematical and
modeling methods. The mathematical method is based on constructing and solving the steady
state equation of the system using the Markov birth and death process for Markov chains or one
type of process [16]. But this process is a complex and separate research work, which is being
carried out in our next studies. In this research work, solutions and results of this problem based
on physical modeling are given.
Based on the model shown in Figure 2 above, management of incoming requests to the server and
priority service during traffic jams were implemented using the system's web server.In this case,
the number of service threads of the web server was evaluated by evaluating the server's ability to
provide service at a certain time (in the study, the server's capabilities in 1 second). The research
was conducted on a server device with technical indicators Montage Jintide® C5220R 2.20 GHz,
RAM 32.0 GB, HDD 500 GB. Also, openresty-1.25.3.1 version of nginx was used as the system
web server.
5. RESULT AND DISCUSSION
The service capabilities of the server in one second were determined based on its technical
indicators according to the methods carried out in the study [3]. In doing so, server process time
(SPT), queuing time (QPT) and server connection time (RTT) were determined under different
system conditions (Table 1). The Apache Jmeter 5.6.3 program was used to generate different
volumes of users and requests in the system for a certain period of time. In this case, it was noted
that the user's internet speed was 23 Mbit/s, and the server's internet speed was 170 Mbit/s.
Table 1. Server Capability Analysis Table
Test
number
Number of
requests in
the system
(from
different
users)
Server
process
time (SPT),
ms
Queuing
time (QPT),
ms
Server
connection time
(RTT), ms
Server
availability, (in
one second)
1
Regular work
times
90 1,49 43
177
2 10 260 1,8 78 77
3 50 465 4,47 102 49
4 100 684 1,48 194 30
5 200 1140 1,5 204 21
6 400 4200 2,43 256 7
7 500 6600 1,86 312 5
8 700 8720 3,35 409 4
9 800 9750 2,05 564 3
10 1000 10035 2,48 632 3
Calculations based on these obtained values estimated the server's service capacity per second at
177. However, when the number of users utilising the system hit 1000, this value decreased to 3.
Because, in this process, the large number of service lines created by the web server for each user
caused the server's response time to increase. This in itself had an impact on the efficiency of the
server.

17
As defined in the research work, the web server's service method for users is adapted to the
model presented in Figure 2, and research on the organization of the service process and the
reduction of the server's loss of requests during times of high traffic of requests were carried out
as follows.
First, the user service algorithm of the nginx web server was analyzed and changes were made to
its operation algorithm.These changes include constantly monitoring whether a service thread is
in use or not, and once the server reaches a concurrent no-rejects service value, users are served
based on whether or not they have previously used the system (Figures 4 and 5, changes and
additions are shown in yellow in Figure 5).
Figure 4. The default service algorithm of the web server for requests
Begin
Begin
Check virtual
server listen port
Check virtual
server listen port
Find the
location
Find the
location
Host=80
Find the
location
Find the
location
Host=443
Read requests
from queue
Read requests
from queue
Find the
location
Find the
location
If other one
Worker
process
event cycle
Worker
process
event cycle
Check is there empty
worker connection
Check is there empty
worker connection
Set response
code to
requests, collect
requests
response data
and sent to
client
Set response
code to
requests, collect
requests
response data
and sent to
client
No
Put requests to
thread pool task
queue
Put requests to
thread pool task
queue
Get request
form task queue
Get request
form task queue
Service the
request
Service the
request
Set keepalive
timeout to
worker
connection
thread
Set keepalive
timeout to
worker
connection
thread
The end
The end

18
Figure 5. Algorithm of the web server for serving requests with priority
The part of the algorithm changes to control the number of threads in constant use served to adapt
the service threads created in the system to the ability of the server to provide services at the
same time. Because, according to the standard settings, the web server creates a separate service
thread for a certain time for any new user admitted to the system.This increased the loading
process of the system and the time it takes to alternately refer to threads in use. As a result,
interruptions in the system's response to temporary requests occurred and led to a sharp decrease
in the server's ability to provide concurrent services.
Therefore, system service threads are assigned to the level of server availability, to ensure the
continuous performance of the server during times of high request traffic and to prevent service
times from exceeding dramatically.
Service requests through the algorithm based on the above model, during high traffic times of
requests, serving requests of users with priority status using the system and temporarily limiting
service requests of new users who have entered the system at that time made it possible. This, in
turn, raised the number of requests handled by an average of 5-7% while maintaining request
service time during peak times nearly unchanged as compared to typical algorithm-based web
Begin
Begin
Check virtual
server listen port
Check virtual
server listen port
Find the
location
Find the
location
Host=80
Find the
location
Find the
location
Host=443
Read requests
from queue
Read requests
from queue
Find the
location
Find the
location
If other one
Worker
process
event cycle
Worker
process
event cycle
Checking the number of
active connection to be less
than the limit value
Set response
code to
requests, collect
requests
response data
and sent to
client
Set response
code to
requests, collect
requests
response data
and sent to
client
More or equal
Put requests to
thread pool task
queue
Put requests to
thread pool task
queue
Less
Get request
form task queue
Get request
form task queue
Service the
request
Service the
request
Set keepalive
timeout to
worker
connection
thread
Set keepalive
timeout to
worker
connection
thread
The end
The end
Check user status
is active or new If active user
If new user

19
servers, resulting in significantly lower server downtime (Figures 6 and 7). Also, all users'
requests have the same priority when serving requests with the default and optimized settings that
are presented in [1], [4], [Error! Reference source not found.], [7], [9], [10], [11]. Therefore,
operating based on the standard algorithm, the probability of rejection was almost the same. For
this reason, most of the requests of users who used the system before were not served.
Figure 6. Request service performance during high traffic times
Figure 7. Request service time (ms)
Based on the above model and the modified algorithm, almost 83% of user requests with service
priority were served, and this was 18% better than a web server with standard settings, and 22%
better than those with optimized settings.As a result, the loss rate of new user requests increased
by 9% compared to the indicators of standard and optimized settings, but compared to the
rejection rate of requests in total, it was less than 5-7% (Figure8). There was also a significant
change in the latency of user requests to connect to the server, and its results are shown in Figure
9.
Figure 8. Rejection rates of requests for user statuses
42.70% 44.36%
37.98%
57.30% 55.64%
62.02%
Default setting Optimized result Changed algorithm result
Rejected requests performance Served requests performance
9000
9200
9400
9600
9800
10000
Default setting
Optimized result
Changed
algorithm result
0.00%
10.00%
20.00%
30.00%
40.00%
50.00%
60.00%
active user requests
lost
new user requests
lost
total lost
optimized settings result
default settings result
changed algorithm settings result

20
Figure 9. Average server connection latency
6. CONCLUSION
In remote service systems, users are offered multi-level assistance for their requests. Initially,
users undergo an identification process. However, during periods of congestion, this process can
lead to decreased system performance. During the identification process, the system effectively
ties up service devices for a certain duration, thereby hindering the processing of requests at
subsequent stages. This inefficiency can lead to delays and decreased overall system
performance. According to the model proposed in this study, serving requests during high traffic
times contributed to alleviating request congestion within the system. Consequently, server
connection delays decreased by 6.2% and 11% respectively, in comparison to systems utilizing
standard settings and optimized settings based on conventional work algorithms. Furthermore,
overall server efficiency experienced a noteworthy enhancement, averaging between 5% to 7%.
Concurrently, the rate of successfully servicing requests from users with priority access during
peak hours surged to an impressive 83%, and it was more to 20% than achieved by the default
algorithm.
REFERENCES
[1] Nedelcu, C. (2015). Nginx HTTP Server. Packt Publishing Ltd.
[2] Zakirov, V., and Abdullaev, E. (2024). Enhancing the efficiency of the remote service process. E3S
Web of Conferences, 501, 02006. EDP Sciences.
[3] Abdullaev, E., Zakirov, V., and Shukurov, F. (2023). Assessment of the distance learning server's
operation strategies and service capacity in advance. E3S Web of Conferences, 420, 06016. EDP
Sciences.
[4] Manchanda, P. (2013). Analysis of optimization techniques to improve user response time of web
applications and their implementation for MOODLE. In Advances in Information Technology: 6th
International Conference, IAIT 2013, Bangkok, Thailand, December 12-13, 2013. Proceedings 6
(pp. 150-161). Springer International Publishing.
[5] Mirhosseini, A., Sriraman, A., and Wenisch, T. F. (2019). Enhancing server efficiency in the face of
killer microseconds. In 2019 IEEE International Symposium on High Performance Computer
Architecture (HPCA) (pp. 185-198). IEEE.
[6] Chi, X., Liu, B., Niu, Q., and Wu, Q. (2012). Web load balance and cache optimization design
based nginx under high-concurrency environment. In 2012 Third International Conference on
Digital Manufacturing an Automation (pp. 1029-1032). IEEE.
[7] Nurmukhamedov, T., and Gulyamov, Z. (2024). Simulation modeling of stock management
systems. In Artificial Intelligence, Blockchain, Computing and Security Volume 2 (pp. 706-711).
CRC Press.
[8] Bartenev, V. (2015). Thread Pools in NGINX Boost Performance 9x. Retrieved from
https://www.nginx.com/blog/thread-pools-boost-performance-9x/
6500 7000 7500 8000 8500
Changed
algorithm setting
result
Optimized setting
Default setting
(MS)

21
[9] Wang, J., and Kai, Z. (2021). Performance analysis and optimization of nginx-based web server.
Journal of Physics: Conference Series, 1955(1), 012033. IOP Publishing.
[10] Ma, C., and Chi, Y. (2022). Evaluation test and improvement of load balancing algorithms of nginx.
IEEE Access, 10, 14311-14324.
[11] Kavon, A. (2020). How To Optimize Nginx Configuration. Retrieved from
https://www.digitalocean.com/community/tutorials/how-to-optimize-nginx-configuration
[12] Zakirov, V. (2024). Model of Assessing the Quality Indicators of the Service Process in Transport.
IETE Technical Review, 41(1), 66-72.
[13] Martinez, J., et al. (2020). Exact response time analysis of fixed priority systems based on sporadic
servers. Journal of Systems Architecture, 110, 101836.
[14] Lozhkovsky, A. G. (2012). Theory of queuing in telecommunications: textbook. Odessa: ONAT
named after A.S. Popova.
[15] Kornyshev, Yu. N. and Fan, G. L. (1985). Information distribution theory. Moscow: Radio and
communication.
[16] M. Zukerman. (2013). Introduction to Queueing Theory and Stochastic Teletraffic Models.- City
University of Hong Kong. 304p.
[17] Гнеденко Б.В., Коваленко И.Н. (1987). Введение в теорию массового обслуживания. –М.:
Наука. Гл. ред. Физ.-мат. Лит. – 336c.
[18] Liu, X., Sha, L., Diao, Y., Froehlich, S., Hellerstein, J. L., & Parekh, S. (2003). Online Response
Time Optimization of Apache Web Server. Quality of Service — IWQoS 2003, 461–478.
https://doi.org/10.1007/3-540-44884-5_25
[19] Jugo, I., Kermek, D., Meštrović, A. (2014). Analysis and Evaluation of Web Application
Performance Enhancement Techniques. Web Engineering, 40–56. https://doi.org/10.1007/978-3-
319-08245-5_3
[20] Kirichek, G., Skrupsky, S., Tiahunova, M., &Timenko, A. (2020). Implementation of web system
optimization method. Computer Modeling and Intelligent Systems, 2608, 199–210.
https://doi.org/10.32782/cmis/2608-16
[21] NishaIlame. (2024). Enhancing Web Application Performance through Database Optimization: A
Comprehensive Study. Journal of Artificial Intelligence, V. 1. – Issue 1. – pp. 1-13.
[22] Rathore, N. (2018). Performance of hybrid load balancing algorithm in distributed web server
system. Wireless Personal Communications, 101, 1233-1246.
[23] Dhingra, A., &Sachdeva, M. (2020). Analyzing the Performance of Web-services during Traffic
Anomalies. International Journal of Advanced Computer Science and Applications, 11(6).
[24] Mission, R. (2021). Web Service Performance Analysis using the GTMetrix Web Monitoring Tool.
Journal of Innovative Technology Convergence, 3(1).
[25] Zatwarnicki, K. (2021). Providing predictable quality of service in a cloud-based web system.
Applied Sciences, 11(7), 2896.
[26] Ibrahim, I. M., Ameen, S. Y., Yasin, H. M., Omar, N., Kak, S. F., Rashid, Z. N., Salih, A.A., Nareen
O. M. Salim., Ahmed, D. M. (2021). Web server performance improvement using dynamic load
balancing techniques: A review. system, 19, 21.
[27] Pavic, F., Brkic, L. (2021). Methods of Improving and Optimizing React Web-applications. 2021
44th International Convention on Information, Communication and Electronic Technology
(MIPRO). https://doi.org/10.23919/mipro52101.2021.9596762
[28] Ahmed, W., Wu, Y., &Zheng, W. (2013). Response time based optimal web service selection. IEEE
Transactions on Parallel and distributed systems, 26(2), 551-561.
[29] Mathew, A. (2019). Cybersecurity infrastructure and security automation. AdvComput: Int J (ACIJ),
10(6).
[30] Ruke, M. M., Gore, S., Chavan, S., &Dakhare, B. (2017). Maintaining data integrity for shared data
in cloud. Advanced Computing: An International Journal (ACIJ), 8(1/2).

22
AUTHOR
VakhidZakirovwas born on September 11, 1960, in the Tashkent city of the Republic of
Uzbekistan. He is Candidate of technical Sciences, in 1988 he defended his thesis for the
degree of candidate of technical Sciences on the topic: "Research and development of a
method for optimizing call quality indicators on telephone networks", specialty 05.12.14
- " Networks, communication centers and information distribution systems».
Associate docent, on the recommendation of the Tashkent electrotechnical Institute of
communications, she was approved by the Higher attestation Commission of the
Republic of Uzbekistan with the academic title of associate Professor, specialty code 05.04.01 - "
Telecommunications and computer systems, networks and devices of telecommunications. Distribution of
information", 06.06.1993.
EldorAbdullaevwas born on February 23, 1993, in the Tashkent region of theRepublic
of Uzbekistan. He is PhD researcher of technical Sciences with Tashkent State
Transport University. He defended his thesis for the master’s degree in 2020.He is the
author of more than thirty scientific works. Right now he is working on his PhD
dissertation in the field of improving and estimating technical aspects of remote
service.

MODEL AND ALGORITHM FOR INCREASING THE EFFICIENCY OF REMOTE SERVICE SYSTEMS SERVER

More Related Content

Similar to MODEL AND ALGORITHM FOR INCREASING THE EFFICIENCY OF REMOTE SERVICE SYSTEMS SERVER (20)

More from acijjournal (20)

Recently uploaded (20)

MODEL AND ALGORITHM FOR INCREASING THE EFFICIENCY OF REMOTE SERVICE SYSTEMS SERVER