'
Dao Minh Sang, Cu Duc Toan, Do Cao Minh
ALGORITHMS AND MODELS FOR MATHEMATICALSIMULATION OF NETWORK PERFORMANCE *
Аннотация:
while assessing the computer networks of businesses, organizations, and educational institutions... In fact, there are numerous ways to assess a computer network system's performance for the best and most efficient outcomes. Reliability enables us to offer the best, targeted solutions to raise the standard of business computer networks. The authors of this paper developed scenarios for evaluation through the use of simulation modeling techniques, and they proposed several approaches for assessing computer network performance.
Ключевые слова:
computer network, network performance, math, evaluate, simulation.
DOI 10.24412/2712-8849-2023-1067-669-675
УДК 004
Dao Minh Sang
Department of Information Technology,
Viet Tri University of Industry
(Viet Tri, Vietnam)
Cu Duc Toan
Department of Information Technology,
Viet Tri University of Industry
(Viet Tri, Vietnam)
Kazan National Research Technological University
(Kazan, Russia)
Do Cao Minh
Department of Information Technology,
Viet Tri University of Industry
(Viet Tri, Vietnam)
ALGORITHMS AND MODELS FOR MATHEMATICAL
SIMULATION OF NETWORK PERFORMANCE
Abstract: while assessing the computer networks of businesses, organizations, and educational institutions... In fact, there are numerous ways to assess a computer network system's performance for the best and most efficient outcomes. Reliability enables us to offer the best, targeted solutions to raise the standard of business computer networks. The authors of this paper developed scenarios for evaluation through the use of simulation modeling techniques, and they proposed several approaches for assessing computer network performance.
Keywords: computer network, network performance, math, evaluate, simulation.
Introduction To support the process of planning and constructing computer networks in order to fulfill the defined goals of computer networks, one of the crucial research tasks for computer networks is to assess and evaluate the performance of computer networks. is to improve system dependability and the quality, effectiveness, and utilization of information. In this article, we outline a few techniques for assessing computer network performance and create test cases for simulation-based evaluation.
The effectiveness and capacity of the network system are measured by network performance [1-3]. Determining the actual efficiency and capacity of the network system under various conditions is what it means to evaluate network performance.
The goal of network [4,5] performance evaluation is to forecast the system's level of performance in order to make comparison assessments when deploying system architecture and developing a computer network system that best satisfies optimal performance requirements.
As a result, the performance evaluation network's circumstances are crucial since they have a direct impact on both the network's performance and the outcomes of the declarative system development process. The scenario as stated has a significant impact on the network efficiency value at the measurement location. Clear definitions of input parameters, such as system network nodes, device connections, participating agents, operating protocols, declared applications, execution time, etc., are required in the scenario.
Research method to evaluate network performance
Mathematical method. Use mathematical techniques that have a wide range of applications, such as statistical probability, graphs, planning, flows, etc., to address a variety of network performance issues. The outcomes of this method are still rather limited, which is a drawback when used in practice because it is very challenging to completely characterize the input components for the problem.
Actual measurement method. The real measuring approach typically yields results with an extremely high degree of precision, but investing too much money and spending too much money before knowing the operating outcomes frequently leads in waste, hence it is only used to track network activity.
Simulation method. The simulation approach is thought to be the most efficient for lowering mathematical method mistakes and system investment costs prior to actual measurement. During system creation, it is essential to accurately and realistically describe the characteristics, methods, and connections between the many participants' effects on the system. Therefore, a very accurate scenario description is required for accurate findings.
In this article, we will delve deeper into understanding and applying simulation methods to evaluate network performance over time.
Real-time simulation is exceedingly challenging, hence the most often used technique at the moment is discrete simulation, in which events are precisely defined and have times that are plotted on a time axis. People use two criteria-state space and its evolution over time-to categorize simulation techniques.
Time-based simulation techniques (also known as synchronous simulation) is a simulation technique in which the control loop of the simulation process controls the time variation in predetermined steps. When the control loop starts, time t is incremented by one step from to t + ... The system will then continue to check to see if any event occurred during the period from [t, t + ]. If present, these events will be executed and the system's state will change accordingly to these events. Assume that the order of occurrence of events in this time period is not important and that these events do not depend on each other. The number of events occurring during this period may also change over time. When t exceeds a certain threshold, the simulation will stop.
Figure 1. Block diagram of the time-based simulation method algorithm
Here, we'll implement a time-based simulation method-based system simulation program. Suppose the queue has an arrival rate of λ and a service rate of μ. In this instance, we employ two state variables: NS is the number of requests being served and is the number of requests in the queue. This system simulation program's objective is to generate a list of time intervals and the state variables that exist throughout each interval.
Assuming is a small enough sea, next we have to use the function draw(p) whose value is true with probability p and false with probability 1 - p. As follows:
Pseudocode 1: Queuing process: arrival → queuing
input ()
t:=0
Ns:=0; Nq=0
While (t<tmax)
do
t:=t+
if draw() then Nq+1
if (Ns=1)
then
if draw(µ)
then if Nq>0
then Nq= Nq-1
else Ns:=0
if (Ns:=0 and Nq:>0)
then Ns:=1; Nq-1
writeln(t,Ns,Nq)
end.
Network Simulation Tool - NS2 (NS - Network Simulator) is an object-oriented discrete event-driven network simulation software to evaluate network system performance, written in C++ programming language, and Otcl.
Simulation scenario. Suppose the Model consists of 5 nodes (0, 1, 2, 3, 4). Simplex-link (One-way data transmission link) between node 0 and node 3 (Bandwidth 0.5 mb, delay 100 ms), node 1 and node 3 (Bandwidth 1.5 mb, delay 200 ms), node 2 and node 3 (Bandwidth 1 mb, delay 300 ms), node 3 and node 4 (Bandwidth 0.8 mb, delay 100 ms). Nodes use DropTail queues.
Figure 2. Simulating packets from nodes 0,1,2 to destination node 4 on NS
Simulation results and discussion. Based on the experimental results of the packet transmission process on NS - 2 above, we can visualize the results of the X graph chart with time according to the specific scenario to clearly see the performance of node 0 (Red) network.), node 1 (Green), node 1 (Blue) as follows:
Figure 3. Network performance of nodes at 1s time
With bandwidth shown on the X axis, and time on the Y axis, so at time = 1s node 0 achieves performance of 0.2 mb/s, node 1 achieves performance of 0.25 mb/s, node 2 achieves performance of 0.4 mb/s.
Table 1. Test results
Time (t) |
t=1s |
||
Node |
0 |
1 |
2 |
Achieve performance (mb/s) |
0.2 |
0.25 |
0.4 |
On the basis of researching and building simulation models to evaluate the performance of computer networks, we can apply them to simple network models in practice, as a basis for certain solutions to improve performance. The network system provides better performance to help agencies, organizations, and businesses build a more stable network model.
REFERENCES:
Номер журнала Вестник науки №10 (67) том 5
Ссылка для цитирования:
Dao Minh Sang, Cu Duc Toan, Do Cao Minh ALGORITHMS AND MODELS FOR MATHEMATICALSIMULATION OF NETWORK PERFORMANCE // Вестник науки №10 (67) том 5. С. 669 - 675. 2023 г. ISSN 2712-8849 // Электронный ресурс: https://www.вестник-науки.рф/article/10557 (дата обращения: 19.05.2024 г.)
Вестник науки СМИ ЭЛ № ФС 77 - 84401 © 2023. 16+
*