Network Simulators: OPNET Overview and Examples
Roman Dunaytsev
Department of Communications Engineering Tampere University of Technology [email protected]
November 30, 2010
Outline
1
About OPNET IT Guru Academic Edition OPNET Modeler Simulation workow Example OPNET Modeler editors
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
2 / 69
Outline
1
About OPNET IT Guru Academic Edition OPNET Modeler Simulation workow Example OPNET Modeler editors
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
3 / 69
About OPNET
Alain Cohen, a 20-year-old MIT student, developed OPNET in 1986
Alain Cohen and his classmate Steven Baraniuk developed a prototype data network modeling and simulation system they called Optimized Network Engineering Tools, or OPNET for short
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
4 / 69
About OPNET (contd)
Alain Cohen, along with his brother Marc and Steven Baraniuk, founded MIL 3, Inc. in 1986 (OPNET 1.1) In 2000, MIL 3, Inc. changed name to OPNET Technologies, Inc. (OPNET 7.0)
Today, OPNET Technologies, Inc. is a provider of software products and related services for:
Application performance management Network planning and engineering Network research and development
www.opnet.com
Chairman & Chief Executive Ocer: Marc Cohen President & Chief Technology Ocer: Alain Cohen
Roman Dunaytsev (TUT) TLT-2707, Guest Lecture November 30, 2010 5 / 69
About OPNET (contd)
Selected consolidated nancial data:
Huge investments in R&D
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
6 / 69
About OPNET (contd)
The companys rst product was OPNET Modeler , a software tool for network simulation and modeling Since then, it has been diversied to provide a range of solutions for: Application performance management
ACE Analyst Standard, ACE Analyst Plus, ACE Enterprise Management Server, OPNET Panorama, ACE Live Appliance, ACE Live Rover, ACE Live on RSP, ACE Live VMon, IT Guru Systems Planner
Network planning and engineering
IT Guru, SP Guru Network Planner, SP Guru Transport Planner, IT Guru Network Planner, NetOne, VNE Server, Report Server, IT NetMapper, IT Sentinel, SP Sentinel, OPNET nCompass for Enterprises, OPNET nCompass for Service Providers
Network R&D
OPNET Modeler, OPNET Modeler Wireless Suite, OPNET Modeler Wireless Suite for Defense
Roman Dunaytsev (TUT) TLT-2707, Guest Lecture November 30, 2010 7 / 69
About OPNET (contd)
Sample list of clients: Service providers
British Telecom, Deutsche Telekom, France Telecom, Inmarsat, . . .
Enterprises
Deutsche Post AG, FBI, Oracle, 20th Century Fox, Xerox, . . .
Network equipment manufacturers
3Com Corporation, Cisco Systems, Ericsson, Fujitsu, HP, Nokia, . . .
Defense and homeland security
Generic Systems Sweden Ab, NASA, US Air Force, US Army, . . .
University Program
Over 25,000 university professors and students use OPNET products in Electrical Engineering, Computer Science, and related disciplines
Roman Dunaytsev (TUT) TLT-2707, Guest Lecture November 30, 2010 8 / 69
About OPNET (contd)
The following products are available through the University Program:
1
IT Guru
Modeling of a broad range of network protocols and technologies 800+ protocol and vendor device models
OPNET Modeler
IT Guru with source code for protocol and technology models
OPNET Modeler Wireless Suite
OPNET Modeler with a broad range of wireless models
SP Guru Transport Planner
Optical network planning and engineering
IT Guru Academic Edition
Based on IT Guru commercial version 9.1 (Build 1999) Created for introductory level networking courses Greatly simplied licensing (6-month renewable license)
Roman Dunaytsev (TUT) TLT-2707, Guest Lecture November 30, 2010 9 / 69
Outline
1
About OPNET IT Guru Academic Edition OPNET Modeler Simulation workow Example OPNET Modeler editors
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
10 / 69
IT Guru Academic Edition
IT Guru Academic Edition limitations:
Limited import and export capabilities Limited wireless functionality Other product modules are not supported The maximum number of simulation events is limited by 50 million The maximum number of intermediate nodes is limited by 20
Supported platforms:
Microsoft: Windows 2000, Windows XP, Windows Vista, Windows 7
ITG Academic Edition v1999.exe: 190 MB
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
11 / 69
IT Guru Academic Edition (contd)
1
Registration
www.opnet.com/university program/itguru academic edition/
Download and installation
www.opnet.com/itguru-academic/download.html
Activation
www.opnet.com/itguru-academic/instructions.html
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
12 / 69
Outline
1
About OPNET IT Guru Academic Edition OPNET Modeler Simulation workow Example OPNET Modeler editors
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
13 / 69
System Requirements
Supported platforms:
Microsoft: Windows 2000, Windows Server 2003, Windows XP, Windows Vista, Windows 7 Linux: Red Hat Enterprise, Fedora
OPNET software does not work with number representation dierent from English
The reason is the dierent decimal separator: point in English, comma practically in all others Start Control Panel Regional and Language Options Standards and formats English (United States)
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
14 / 69
Installing OPNET Modeler
Supporting software for OPNET Modeler:
Microsoft: Visual C/C++ 6.x, Visual Studio .NET 2002, Visual Studio .NET 2003, Visual Studio 2005, Visual Studio 2008, or Visual C++ 2008 Express Edition Linux: gcc 3.4 or higher
Install OPNET components in the following order:
1 2 3 4
Software Additional modules (if any) Model library Documentation
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
15 / 69
Accelerating Network R&D
OPNET network R&D solutions enable:
Test technology designs in realistic scenarios Evaluate enhancements to standards-based protocols Develop new protocols and technologies
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
16 / 69
Key Features
Hierarchical GUI-based editors High-delity modeling
800+ wired/wireless protocol and vendor device models with source code Dierent aspects of wireless communication, including RF propagation, antenna modeling, signal modulation, node mobility, and interference
Scalable simulation
32-bit and 64-bit fully parallel simulation kernel Grid computing support for distributed simulation
Sophisticated analysis
Integrated GUI-based debugging and analysis
Integrating live network and application behavior
Optional System-in-the-Loop module to interface simulations with live systems Open interface for integrating external les, libraries, and other simulators
Roman Dunaytsev (TUT) TLT-2707, Guest Lecture November 30, 2010 17 / 69
Simulation Technologies
OPNET supports 4 simulation technologies:
Discrete Event Simulation (DES) Flow Analysis ACE QuickPredict Hybrid Simulation (within the DES environment)
OPNET simulation technologies
Discrete Event Simulation (DES)
Flow Analysis
ACE QuickPredict
Hybrid
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
18 / 69
Simulation Technologies (contd)
Discrete Event Simulation provides highly detailed models that explicitly simulate packets and protocol messages The models in DES execute the protocol in much the same way as a production environment Although DES provides very high-delity results, simulation runtimes are longer than with the other methods
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
19 / 69
Simulation Technologies (contd)
Hybrid simulation combines 2 distinct modeling techniques (analytical and discrete) to provide accurate, detailed results for targeted ows Hybrid simulation relies on background and explicit trac
Background trac is used to represent most of a networks ambient load at an abstract level Selected network application ows are represented in detail, using the explicit trac models
Execution runtimes can be signicantly faster as compared with DES
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
20 / 69
Simulation Technologies (contd)
Flow Analysis uses analytical techniques and algorithms to model steady-state network behavior Flow Analysis does not model individual protocol messages or packets, therefore it does not generate results for transient network conditions It can be used to study routing and reachability across the network in steady state, and in scenarios with one or more failed devices Execution runtimes can be signicantly faster as compared with DES
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
21 / 69
Simulation Technologies (contd)
ACE QuickPredict uses an analytical technique for studying the impact on application response time of changing network parameters (e.g., bandwidth, latency, utilization, packet loss)
This technique is supported within the OPNET Application Characterization Environment (ACE)
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
22 / 69
Outline
1
About OPNET IT Guru Academic Edition OPNET Modeler Simulation workow Example OPNET Modeler editors
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
23 / 69
Projects and Scenarios
OPNET Modeler uses a project-and-scenario approach to model networks Project a collection of network-related scenarios, each of which explores a particular aspect of the network design
All projects contain at least 1 scenario
Scenario a single instance of a network
Typically, a scenario presents a unique conguration for the network The term conguration can refer to dierent aspects such as topology, protocols, applications, trac, and simulation settings
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
24 / 69
Projects and Scenarios (contd)
OPNET simulation workow :
1 2
Create a project Create a baseline scenario
Import or create a network topology Import or create trac Choose statistics to be collected Run the simulation View the results
Duplicate the scenario
Make changes Re-run the simulation Compare the obtained results
Repeat Step 3 if needed
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
25 / 69
Projects and Scenarios (contd)
New project
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
26 / 69
Projects and Scenarios (contd)
Project Editor is used to construct and edit the topology of a network model
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
27 / 69
Projects and Scenarios (contd)
Project Editor window
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
28 / 69
Network Topologies
Initial topology
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
29 / 69
Network Topologies (contd)
There are several methods for creating a network topology:
Manually, by dragging and dropping objects from an Object Palette to the Project Editor workspace Manually, using the Topology Rapid Conguration... command from the Menu Bar to specify and build a complete network topology quickly Automatically, by importing the network model from an external data source either a system that monitors your network or one or more data les that describe the network
Importing a topology ensures that the network model you build corresponds to the existing network exactly
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
30 / 69
Network Topologies (contd)
Network scale
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
31 / 69
Network Topologies (contd)
Background maps
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
32 / 69
Network Topologies (contd)
Zooming
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
33 / 69
Network Topologies (contd)
Dragging and dropping objects from an Object Palette into the Project Editor workspace
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
34 / 69
Network Topologies (contd)
Using the Topology Rapid Conguration... command from the Menu Bar to quickly deploy common network topologies
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
35 / 69
Network Topologies (contd)
Available congurations: Bus, Mesh (Full or Randomized), Ring, Star, Tree, and Unconnected Net
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
36 / 69
Network Topologies (contd)
Using the Topology Deploy Wireless Network... command from the Menu Bar to specify and build a wireless network
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
37 / 69
Model Library
OPNET Modeler provides an extensive library of models that you can use to build networks
www.opnet.com/support/des model library/index.html
These models are called standard models because users can also develop their own models
Those models can then be shared with other OPNET users if desired
Certain models support the needs of users with particular interests in emerging or vendor-specic technologies (aka specialized models)
An additional license is needed to use these models in a simulation
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
38 / 69
Model Library (contd)
The standard model library consists of the following types of objects:
Subnetworks Nodes (aka devices) Links LANs and clouds Utility objects
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
39 / 69
Objects
Model Family: internet toolbox
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
40 / 69
Objects (contd)
Subnetworks
Subnetworks are essentially containers that abstract the network components specied within them into one object A subnetwork can also contain other subnetworks A special subnetwork called the top level or global subnetwork is the highest level subnetwork in the network hierarchy
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
41 / 69
Objects (contd)
Nodes
A node represents a network device with a wide range of possible capabilities (router, switch, hub, workstation, server, rewall, etc.) The actual function and behavior of a node is determined by its node model
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
42 / 69
Objects (contd)
Links
Links represent the physical media and properties (line rate in bits per second, delay, likelihood of data corruption, etc.) Links are represented as line segments or a series of line segments with arrowheads
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
43 / 69
Objects (contd)
LANs
A LAN object abstracts the LAN infrastructure into one object LAN objects dramatically reduce the amount of conguration required to do to represent an internetwork of LANs, and the amount of memory needed to run the simulation
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
44 / 69
Objects (contd)
Clouds
A cloud object abstracts the WAN infrastructure into one object Cloud objects provide high-level characteristics (packet latency and discard ratio) used to simulate the behavior of ATM, Frame Relay, and IP WANs
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
45 / 69
Objects (contd)
Utility objects
Utility objects do not correspond to the actual physical infrastructure Instead, they perform logical functions in the network (conguration of network resources, scheduling special events, etc.)
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
46 / 69
Applications and Trac
The rst step is to drag and drop Application Cong and Prole Cong objects from the Object Palette to the Project Editor workspace Application Cong species standard and custom applications used in the simulation, including trac and QoS parameters
Standard applications (Light/Heavy): Database, Email, FTP, HTTP, Print, Remote Login, Video Conferencing, Voice
Prole Cong species the activity patterns of a user or groups of users in terms of the applications used over a period of time
You can have several dierent proles running on a given workstation or a LAN These proles can represent dierent user groups and behavior patterns
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
47 / 69
Applications and Trac (contd)
LAN users (1 + 2 + 7 = 10) Profiles (3) Apps (6) Professor (1) Secretaries (2) DB access HTTP light Researchers (7) HTTP heavy
VoIP
Email
Email
FTP
Proles describes activity patterns, such as:
When does a user start using applications? What is the duration of his/her activity? What applications does he/she use? How often does he/she use each application?
Congure applications Dene proles (Menu Bar Protocols Applications Deploy Dened Applications...)
Roman Dunaytsev (TUT) TLT-2707, Guest Lecture November 30, 2010 48 / 69
Choosing Statistics
Choose statistics to collect
Menu Bar DES Choose Individual Statistics... Or Right-click in the Project Editor Choose Individual DES Statistics List of statistics appears
Types of statistics
Global: collected on the total network (e.g., application response time) Node: collected on individual nodes (e.g., delay, delay variation) Link: collected on individual links (e.g., utilization, throughput, queuing delay)
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
49 / 69
Choosing Statistics (contd)
Choose Results dialog box
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
50 / 69
Running Simulation
Menu Bar DES Congure/Run Discrete Event Simulation... Set simulation options and click Run
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
51 / 69
Viewing Results
Menu Bar DES Results View Results... Or Right-click in the Project Editor View Results
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
52 / 69
Outline
1
About OPNET IT Guru Academic Edition OPNET Modeler Simulation workow Example OPNET Modeler editors
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
53 / 69
Small Internetworks
Case study: Small Internetworks In this example, you plan for the expansion of a small companys intranet Currently, the company has a star topology network on the rst oor of its oce building and plans to add an additional star topology network on another oor You will build and test this what-if scenario to ensure that the load added by the second network will not cause the network to fail
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
54 / 69
Small Internetworks (contd)
Creating the network Initial Topology: Create empty scenario Choose Network Scale: Oce & Use metric units Specify Size: 100 m x 100 m Select Technologies: Sm Int Model List
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
55 / 69
Small Internetworks (contd)
Rapid Conguration: Star Center Node Model: 3C SSII 1100 3300 4s ae52 e48 ge3 Periphery Node Model: Sm Int wkstn Number (of periphery nodes): 30 Link Model: 10BaseT Center X x Y: 25 x 25 Radius: 20 Sm Int server, 10BaseT, Sm Application Cong, Sm Prole Cong
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
56 / 69
Small Internetworks (contd)
Right-click on a node View Node Description 3C SSII 1100 3300 4s ae52 e48 ge3 represents a stack of 4 3Com switches (4s):
2 SuperStack II 1100 switches 2 SuperStack II 3300 switches
52 auto-sensing Ethernet ports (ae52) 48 Ethernet ports (e48) 3 Gigabit Ethernet ports (ge3)
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
57 / 69
Small Internetworks (contd)
Compare with an abstract node in ns-2
set node 30 [$ns node]
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
58 / 69
Small Internetworks (contd)
The original network
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
59 / 69
Small Internetworks (contd)
Will the server be able to handle the additional load of the second network?
Right-click on the server node Individual DES Statistics Node Statistics Ethernet Load (bits/sec)
Will the total delay across the network be acceptable once the second network is installed?
Right-click in the workspace (but not on an object) Individual DES Statistics Global Statistics Ethernet Delay (sec)
Run the simulation for 30 minutes
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
60 / 69
Small Internetworks (contd)
Expanding the network Rapid Conguration: Star Center Node Model: 3C SSII 1100 3300 4s ae52 e48 ge3 Periphery Node Model: Sm Int wkstn Number (of periphery nodes): 15 Link Model: 10BaseT Center X x Y: 75 x 62.5 Radius: 20 CS 2514 1s e2 sl2 (Cisco 2514 router), 10BaseT
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
61 / 69
Small Internetworks (contd)
The extanded network
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
62 / 69
Small Internetworks (contd)
Comparing results Menu Bar DES Results Compare Results...
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
63 / 69
Small Internetworks (contd)
The average load for the expansion scenario is higher (as expected) But there is no signicant change in Ethernet delay on the network
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
64 / 69
Outline
1
About OPNET IT Guru Academic Edition OPNET Modeler Simulation workow Example OPNET Modeler editors
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
65 / 69
OPNET Modeler Editors
The Project Editor is used to construct and edit the topology of a network model
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
66 / 69
OPNET Modeler Editors (contd)
The Node Editor provides operations to support creation and editing of node models
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
67 / 69
OPNET Modeler Editors (contd)
The Process Editor is used to specify the behavior of process models
Process models use a nite state machine (FSM) paradigm to express behavior that depends on the current state and a new stimuli
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
68 / 69
OPNET Modeler Editors (contd)
The operations performed by a process model are described in statements based on the C or C++ languages These statements can be associated with states, transitions, or special blocks within the process model
Roman Dunaytsev (TUT)
TLT-2707, Guest Lecture
November 30, 2010
69 / 69
