[Switch] comparison of several Network Simulation Software

Transferred from:

Network simulation is a simulation technology that establishes statistical models of network devices and network links and simulates the transmission of network traffic to obtain network performance data required for network design or optimization. Because the simulation is not based on mathematical computing, but on a statistical model, the randomness of statistical reuse is precisely reproduced.
Network Simulation Technology has the following features: 1. The new simulation experiment mechanism gives it the ability to obtain highly reliable results in a highly complex network environment. 2. The prediction function of network simulation is incomparable to any other method. 3. It can be used for optimization and expansion of existing networks or design of new networks, it is particularly suitable for the design and optimization of medium and large networks. 4. The initial application cost is not high, and the established network model can be used continuously, and the investment will continue to decrease in the future.

OPNET

Product performance Overview: OPNET is designed for professionals to help customers design, construct, analyze, and manage network structures, devices, and applications. OPNET products mainly target three types of customers and are divided into four series. Three types of customers refer to network service providers, network equipment manufacturers, and general enterprises.

The core products of the Four series include: 1. serviceproviderguru: intelligent network management software for network service providers. Is the latest product of OPNET. 2. OPNET Modeler: provides a network technology and product development platform for technicians (engineers. It helps them design and analyze networks, network devices, and communication protocols. 3. itguru: Helps network professionals predict and analyze the performance of network and network applications, diagnose problems, find bottlenecks affecting system performance, and propose and verify solutions. 4. WDM guru, used for analysis and evaluation of wavelength division multiplexing optical fiber networks.

I used OPNET modeler8.1, So I compared it with other network simulation software. OPNET technology's simulation software OPNET has the following outstanding features, so that it can meet the simulation needs of large-scale complex networks:

1. the layer-3 modeling mechanism is provided. The bottom layer is the process model, and the protocol is described by the state machine. The node model is composed of the corresponding protocol model, reflecting the device characteristics. The top layer is the network model. The layer-3 model corresponds to the actual network, device, and protocol layers, fully reflecting the characteristics of the network;

2. A complete basic model library is provided, including routers, switches, servers, clients, ATM devices, DSL devices, and ISDN devices;

3. The Discrete Event driver is used to improve the computing efficiency compared with the time driver.

4. The hybrid modeling mechanism is used to combine the packet-based analysis method with the statistical-based mathematical modeling method, which can not only obtain detailed simulation results, but also greatly improve the simulation efficiency.

5. OPNET has rich functions for statistic collection and analysis. It can directly collect common performance statistics parameters at various network layers, and conveniently compile and output simulation reports.

6. Provides interfaces with the network management system and traffic monitoring system, which can easily use the existing topology and traffic data to establish a simulation model. At the same time, the simulation results can be verified.

Disadvantages of OPNET:

1. the expensive. OPNET license fee for a single user exceeds 20 thousand US $. Piracy is rampant.

2. the barriers to entry into learning are high. It takes at least one month to become more proficient through specialized training.

3. When the simulation network scale and traffic are large, the simulation efficiency will be reduced. Current solution: hierarchical modeling is used to converge network traffic and simplify the network model. The background traffic and foreground traffic are combined. The traffic proportion compression method is used to optimize and adjust the simulation parameter design; simplified routing traffic, result analysis, and selection of appropriate result collection and processing methods based on different statistical parameters. Source: (Guangdong Telecom Research Institute of Science and Technology)

4. The model library provided by the software is limited. Therefore, the modeling of some special network devices must be implemented by programming nodes and process layers. (. The standard result parameters provided by the network simulation software cannot meet all the requirements of the actual user. If you need to collect some special parameters of the network device, you must program at the process level to collect network parameters that you are interested in. B. in general, the network protocol models provided by manufacturers all lag behind the standard publication days of months or even one year, you only need to develop your own protocol model through programming. C. for large-scale network simulation, sometimes it is necessary to change some features of the model through programming to improve the simulation computing efficiency according to the actual situation .) Network Element Modeling involving underlying programming is technically difficult, because you need to have a deep understanding of protocols, standards, and implementation details, and master the complex Modeling Mechanism of network simulation software. Therefore, professional and technical personnel who have been specially trained are generally required to complete the process. The difficulty of programming limits the popularity and promotion of OPNET. In addition, the simulation platform built on OPNET cannot be separated from the OPNET environment, which is also a limitation of OPNET.

5. OPNET is suitable for the simulation of the routing protocol, but the simulation of the link can only be done through pipeline stage. For example, a user tried to change the transmission speed of the radio propagation delay model in the pipeline stage (the default value is the speed of light) to the speed of the acoustic wave in the sea by 1500 m/s, however, the simulated propagation delay results have not changed much. Based on experience, MATLAB or spw is more suitable for link layer simulation.

Introduction to NS (Network Simulator)

NS is an excellent simulation software developed by UC Berkeley to simulate various IP networks. The software was initially developed for Network Design and Simulation Based on UNIX systems.

TCL (tool command language) and TK are two packages installed in Unix/Linux environment. Together, they constitute a set of development system applications and graphical user interface (GUI) application environment. TK is an extension of TCL in the X Window environment. It contains all the C library functions of TCL and C library functions such as windows and controls supporting X Windows, it provides convenience for users to develop graphical user interfaces.

The NS design uses two programming languages: C ++ and OTCL. Both programming languages are object-oriented. The C ++ program module runs very fast and is a mandatory programming language that is easy to implement precise and complex algorithms, but can be modified and discovered, fixing a Bug takes a long time because it is complicated. OTCL is a scripting language that does not have a mandatory type. It is relatively simple, easy to implement and modify, and easy to detect bugs, although its running speed is much slower than that of the C ++ module.

Simulation Principle of NS-network components. Nsobject is the parent class of all basic network components. Its parent class is tclobject. The object of this class has a basic function, that is, to process packets (packet ). All basic network components can be divided into Classifier and connector ). They are both nsobject and direct subclasses, and are also the parent class of all basic network components. The Component Objects of the derived class of the classifier include the address classifier and multicast classifier. The derived class Component Objects of the connector include queue, latency, various proxies, and tracing object classes. An application is a simulation of an application built on a transport proxy. There are two types of "Applications", data source generators and simulated applications in NS2. NS is a discrete event-driven network simulator. It uses event scheduler to list and maintain the work that all components want to complete and the time when the work is scheduled.

The NS operating platform can be windows, Linux, Uinx, and machitosh. a c ++ compiler must be installed in the system. NS workflow: NS code is written in the OTCL language, interpreted by the OTCL interpreter, compiled and simulated using the NS simulation library, output the simulation results, and recorded Based on the simulation results, you can further analyze the relevant content to generate a network extension diagram or a visual chart of the data. By using the auxiliary Nam tool, you can clearly display the network extension diagram in NS. Using the X graph tool, you can display the simulation results of ns in charts,

The starting point of the NS design is network simulation. It integrates multiple network protocols, business types, routing Queuing management mechanisms, and routing algorithms. In addition, NS integrates multicast services and MAC layer protocols related to LAN simulation. The simulation mainly targets the routing layer, transmission layer, and data link layer. Therefore, the NS can simulate fixed, wireless, satellite, and hybrid networks. However, it is most suitable for TCP layer and above simulation. NS features open source code, high scalability, and obvious speed and efficiency advantages.

Advantages and disadvantages of NS and OPNET:

1. The relationship between OPNET and NS2. They are also the best match for the installation and use in these two environments (OPNET in win and NS2.

2. OPNET is a commercial software, so the interface is very good. The function is very powerful, the interface is incorrect, unified and strict. The operation is also very convenient, and the main modification to the node is to modify its attributes. It is both an advantage and a disadvantage. If you need a special node or a node that is suitable for you, it is not as convenient as NS2. At the same time, the release of the commercial software version is not as fast as that of the NS2.

3. NS2. it is free software, which is the biggest advantage compared with OPNET. Therefore, it enjoys a high popularity and is a strong competitor of OPNET.

4. The ns2-based interface is not as good as OPNET. Although its functions are powerful, there are a lot of scattered things. It is very troublesome and difficult to get started when I first learned it. The operation is convenient, but because it is not developed by the same company, the format is not very uniform, and the instruction manual is not well organized. There is no ready-made node or something you can use C ++ to compile your own code, and you can construct the node you want based on your own intentions.

MATLAB software

MATLAB is a scientific computing system environment launched by Mathworks for numerical computing and graphic processing. MATLAB is the abbreviation of matrix lab. In the MATLAB environment, you can perform various operations such as program design, numerical calculation, drawing, input/output, and file management. MATLAB provides a mathematical system environment for human-machine interaction. The basic data structure of the system is a matrix. Explicit dimension is not required when a rectangular object is generated. Compared with programming using C or FORTRAN for numerical computation, MATLAB can save a lot of programming time.

MATLAB's good openness and operational reliability have led to the elimination of closed software packages (such as UMIST in the UK, Lund in Sweden and keddc in simnon and Germany) in the field of control, it was rebuilt on the MATLAB platform.

In, Mathworks released version, 6. while inheriting and developing its original numerical computing and graphical visualization capabilities, Version X has made the following important changes. This is an interactive dynamic system modeling, simulation, and analysis integrated environment. Its appearance makes it possible for people to consider many non-linear and random factors that had to simplify assumptions before, thus greatly improving people's cognitive ability to nonlinear and Random Dynamic Systems. (2) developed components for direct data exchange with the outside, and opened up the path of MATLAB for real-time data analysis, processing and hardware development. (3) A symbolic computing toolkit is released. In 1993, Mathworks acquired the right to use maple from the University of Waterloo, Canada, and developed symbolic math toolbox 1.0 with Maple as the "engine. The move of Mathworks accelerates the conclusion of a long-standing debate on numerical computation and symbolic computation, leading to a new era of complementary development of the two computation methods. (4) create a notebook. Mathworks aims at the most widely used word and uses DDE and OLE to achieve seamless connection between MATLAB and word, as a result, professional science and technology workers have created a high level of integrated computing, graphic visualization, and text processing environment.

The mtalab system consists of five main parts.

(1) matalb language system matlab is a high-level matrix/array language with features such as conditional control, function calling, data structure, input and output, and object-oriented programming language. It can be used to complete basic tasks of algorithm design and algorithm experiment, as well as large-scale programming to develop complex applications.

(2) MATLAB work environment is a general term for the management functions provided to users by Ma Ding lab. it includes methods and methods for managing input and output of variables in a workspace, and various tools for developing, debugging, and managing M files.

(3) The graphic sentence system is the basis of the MATLAB graphics system, including high-level MATLAB commands to Complete 2D and 3D data diagrams, image processing, Animation Generation, graphic display, and other functions, it also includes low-level MATLAB commands for controlling features of graphical images and other objects, as well as various tools for developing GUI applications.

(4) MATLAB mathematical function library is a general term for various mathematical algorithms used by MATLAB. It includes various elementary function algorithms, including matrix operations, matrix analysis, and other high-level mathematical algorithms.

(5) Matlab application interface (API) is a function library provided by MATLAB for users to use C programs or fortran programs in the MATLAB environment, including the function of calling the program (Dynamic Link) from MATLAB to read and write mat files. It can be seen that MATLAB is a powerful system and an Environment Integrating numerical computing, graphic management, and program development. In addition, Ma Ding lab has a powerful function expansion capability. Along with its main system, it can be equipped with a variety of toolboxes to complete some specific tasks.

In international academic circles, MATLAB has been recognized as an accurate and reliable scientific computing standard software. MATLAB applications can be seen in many world-class academic publications (especially information science publications. In design research units and industrial sectors, Matlab is considered as the preferred software tool for efficient research and development. Such as the national instruments Signal Measurement and Analysis Software LabVIEW, Cadence signal and communication analysis design software spw, or directly built on Matlab, or Matlab as the main support. Another example is HP's VXI hardware, TM's DSP, and Gage's various hard cards and instruments are supported by Matlab.

Disadvantages of MATLAB:

1. The disadvantage of MATLAB is that it is slower than other advanced programs. Because the MATLAB program does not need to be compiled or preprocessed, and does not generate executable files, the program is interpreted and executed, so the speed is slow.

2. Another disadvantage of MATLAB is that it cannot implement port operations and real-time control. However, combined with the use of C ++ builder, this shortcoming can be overcome by implementing complementary advantages.

3. another disadvantage of MATLAB is that the software is too large. According to the popular version 5.2, it has more than 400 mb and occupies nearly 1 GB of hard disk space. Generally, earlier computers cannot be installed.

Price: RMB 1900 for genuine MATLAB 6.5.1 software is available online.

Other Simulation Software

Cassap Simulation Software

Traditional design methods cannot be used to design a complex and huge system such as spread-spectrum digital mobile communication. Synopsys) the design tool cossap provided by the company helps designers establish a system model at the algorithm concept level, which can define and change system parameters at any time, simulate the system signal processing flow to study the algorithm performance and indicators of the system. Cossap's design philosophy is the top-down (top-down) design method, that is, the system performance analysis, algorithm analysis, and hardware, software division, design, finally, the system design can be completed only after the software and hardware are combined, simulated, commissioning and verification.

Cassap provides a most effective simulation tool for developers of digital signal processing systems, allowing engineers to perform algorithm simulation at the conceptual, algorithm, and architectural layers. Its first data stream driver simulator is 8-16 times faster than the traditional time-cycle-based simulator. More than 1000 high-level modules are provided for Dynamic Data Stream-driven simulation (asynchronous and multi-rate support), and the required modules can be automatically generated with behavior-level or RTL-level VHDL, it is used for ASIC/FPGA implementation and can also generate DSP code of various styles for software implementation by DSP processors. Cassap can be widely used in digital transmission systems that require analysis and evaluation algorithms and implementation methods, such as communication, images, and multimedia, it also provides a dedicated development platform for GSM, CDMA, DECT and other standards.

Spw Simulation Software

Spw simulation software is a product of cadence. It provides a modular design, simulation and implementation environment for electronic systems. It is used for algorithm development, filter design, and C code generation, ideal environment for hardware/software structure joint design and hardware integration.

The most outstanding part of spw is its interface with HDS and Matlab. Many models in MATLAB can be directly transferred to spw, and then use HDS to generate C language simulation code or HDL language simulation code. That is to say, if it is simple, you can directly use MATLAB to create a model, and then implement the layout.

It can be said that spw includes many functions of Matlab, and even the demo is a bit like .. Its common application fields include wireless and wired carrier communications, multimedia and network devices.