Transcript Teaching with OPNET Modeler

Teaching with OPNET
Software
Dr. Vasil Y. Hnatyshin
Department of Computer Science
Rowan University
Copyright © 2005 OPNET Technologies, Inc. Confidential, not for distribution to third parties.
Outline
 Course Overview
 List of Topics and textbooks
 Software Tools and Projects
 OPNET IT Guru and Modeler Laboratories
 Teaching with IT Guru
 Advantages
 Disadvantages
 Alternative IT Guru Lab Manual style
 Idea
 Advantages/disadvantages
 Simple Example
 Teaching with OPNET modeler
 Conclusions
Copyright © 2005 OPNET Technologies, Inc. Confidential, not for distribution to third parties.
2
Course Overview: List of Topics
 “TCP/IP and Internet Technologies”, Spring 2005
 W. Stallings, "Computer Networking with Internet Protocols and Technology"
 List of Topics:
 Introduction:
 OSI and TCP/IP reference models
 Encapsulation and TCP/IP header format
 TCP internals:
 Reliability and Congestion control
 TCP flavors: Tahoe, Reno, New Reno, SACK, Vegas
 IP internals:
 IP addressing, subnets and CIDR
 IPv4, ICMP, and IPv6
 Internet Technologies and Quality of Service:
 Traffic regulation and Token bucket
 Scheduling and Queue Management: FIFO, PQ/TPQ, WRR/WFQ, RED
 Integrated and Differentiated Services
Copyright © 2005 OPNET Technologies, Inc. Confidential, not for distribution to third parties.
3
Course Overview: Tools and Projects
 TCPDump: Study of network
characteristics
 Number of TCP vs UDP packets
 Number of TCP SYN and FIN packets
 Packet Size distribution
 Use of TOS byte
 Use of TCP options
 OPNET Modeler
 Implementation of Bandwidth
Distribution Scheme
Copyright © 2005 OPNET Technologies, Inc. Confidential, not for distribution to third parties.
 OPNET IT Guru:
 TCP vs. UDP: influence of different
transport protocols on application and
network performance
 TCP Flow Control: influence of
window size, MTU, and other
parameters on application and
network performance
 TCP Throughput: influence of window
size, MTU, and other parameters on
application and network performance
 QoS: Impact of Queuing Policy:
influence of different queuing
mechanisms on delay, jitter, and traffic
loss
4
OPNET IT Guru and Modeler
Laboratories
OPNET IT Guru:
 Laboratory manuals by Leann Christianson and Kevin Brown to accompany
Data and Computer Communications and Business Data Communications
by William Stallings
 Lab #2: TCP and UDP
 Lab #4: TCP Parameters for Flow Control
 Lab #7: TCP throughput
 Lab #5: QoS: Impact of Queuing Policy
 OPNET IT Guru Tutorials:
Introduction
TCP Internals
TCP Internals
Technologies & QoS
Introduction to IT Guru
 Introduction
 Small Networks
OPNET Modeler:
 OPNET Modeler Tutorials:
 Small Networks
 Basic Processes
 M/M/1
Copyright © 2005 OPNET Technologies, Inc. Confidential, not for distribution to third parties.
5
Advantages of using IT Guru in the class
 Students understand the course material better:
 Excellent illustration of numerous networking concepts such as:
TCP’s slow start, congestion avoidance, fast recovery/fast
retransmit
 Better understanding of the importance of the protocol
configuration parameters.
 Hands-on experience in configuring and setting up simple
network configurations
 Increased student’s enthusiasm
 Simple to download and install IT Guru software
Copyright © 2005 OPNET Technologies, Inc. Confidential, not for distribution to third parties.
6
Disadvantages of using IT Guru in the
class
 Lack well-fitting laboratory manuals*
 The laboratory manuals never follow the course outline
 Laboratory assignments often confusing
 Difficult to achieve the same results as those shown in the manual
 Different versions of the ever-changing software
 Following laboratory instructions step-by-step does not facilitate learning
 Initially:
 Difficult to explain how to use IT Guru to the students
 Significantly more work for the instructor 
*Based on the Laboratory manuals by Leann Christianson and Kevin Brown to accompany Data and
Computer Communications and Business Data Communications by William Stallings.
Copyright © 2005 OPNET Technologies, Inc. Confidential, not for distribution to third parties.
7
Disadvantages of using IT Guru in the
class
Available OPNET IT Guru laboratory manuals
 Data and Computer Communications by W. Stallings, manual by L.
Christianson and K. Brown
 Business Data Communications by W. Stallings, manual by L. Christianson
and K. Brown
 Business Data Networks and Communications by R. Panko
 Computer Networks – A Systems Approach by Peterson & Davie,
 IT Guru tutorials
 OPNETWORK tutorials on the web
But the topics covered in class vary from instructor to
instructor and almost always requires more.
Copyright © 2005 OPNET Technologies, Inc. Confidential, not for distribution to third parties.
8
Disadvantages of using IT Guru in the
class
Laboratory assignments often too simple and too confusing, why?
Laboratory assignments consists of


Step-by-step configuration instructions
Questions regarding collected results
Problems with this approach:




Requires no thinking during configuration part
Lab manual’s IT Guru version may be different from actual software
Easy to make a mistake
Simulation results often incorrect or different from those in the manual
Copyright © 2005 OPNET Technologies, Inc. Confidential, not for distribution to third parties.
9
Alternative Lab Manual style
 Have two sets Laboratory manuals
 IT Guru tutorials that introduce students to software basics
 Actual laboratory assignments
 The laboratory assignments would consists of
 Topology/Applications/Protocols configuration specifications
 No explicit step-by-step instructions
 Questions about collected results
 Students will submit:
 Summary of configuration steps
 Answers to the question
 Laboratory project files
Copyright © 2005 OPNET Technologies, Inc. Confidential, not for distribution to third parties.
10
Alternative Lab Manual style
Advantages
 Will require student thinking during configuration phase
 Configuration errors and incorrect results are easier to track as
the students provide the summary of their configuration steps
 Easier to write and change lab manuals
 Software updates have less effect on the labs
 More work for the students
 Hopefully less work for the instructor
 Disadvantages
 May be difficult for students to learn how to use IT Guru
 More work for the students
Copyright © 2005 OPNET Technologies, Inc. Confidential, not for distribution to third parties.
11
Alternative Lab Manual style: Example
Lab: TCP Flow Control – Topology configuration
Node E-mail Client sends e-mail traffic to node E-mail Server over an IP network.
The nodes connected to the network via DS 1 link.
Packets traveling through given IP network experience latency which is exponentially
distributed and has mean outcome of 0.25 seconds.
Copyright © 2005 OPNET Technologies, Inc. Confidential, not for distribution to third parties.
12
Alternative Lab Manual style: Example
Lab: TCP Flow Control – Protocol/Application Configuration
E-mail Configuration:
 E-mail size: 1 MByte for all messages
 E-mail interarrival time: exponentially distributed with mean outcome of 100 seconds
 E-mail application always generates messages in groups of 3.
Client’s Application configuration:
 Application Type: E-mail
 Application start time: uniformly distributed between 100 and 110 seconds
 Application end time: End of simulation
 Number of concurrent applications: 1
TCP configuration:
 Maximum Segment Size: 512 bytes
 Receiver Buffer: 8760 bytes
 Other parameters: default values
Copyright © 2005 OPNET Technologies, Inc. Confidential, not for distribution to third parties.
13
Alternative Lab Manual style: Example
Lab: TCP Flow Control – Assignment
 Examine E-mail application performance:




Download Response Time
Upload Response Time
Traffic Sent
Traffic Received
 Examine TCP performance parameters:
 Application data Delay
 Retransmission Count
 Segment Delay
1. Repeat the study with the values of TCP receiver buffer values of 16K, 32K, 64K.
2. Repeat the study with the TCP Max Segment size values of 1024 and 2048 bytes,
Receiver buffer size is set to 32 K.
Plot, compare, and explain obtained results.
Copyright © 2005 OPNET Technologies, Inc. Confidential, not for distribution to third parties.
14
Teaching with OPNET Modeler
 Learning OPNET Modeler
 Tutorials
 Documentation
 Code
 Endless one-on-one sessions
 Outcome:
 Better understanding of the router internals
 Two working implementations of the Bandwidth Distribution Scheme
 A published paper:
 “Influence of the discarding the out-of-profile packets policy on TCP traffic," In
Proceedings of the 2005 International Conference on Parallel and Distributed
Processing Techniques and Applications (PDPTA'05)
 Using OPNET Modeler in the undergraduate course is difficult
 Easier to use it as group research projects such as independent study
and senior project
Copyright © 2005 OPNET Technologies, Inc. Confidential, not for distribution to third parties.
15
Summary and Conclusions
 There is need for additional OPNET IT Guru laboratory Manuals
 Alternative ways to structure and present the IT Guru manuals
 Yet,OPNET IT Guru is an excellent software tool that facilitates the
classroom learning process
 Using OPNET Modeler at the undergraduate level is possible but is
better suited for outside-the-classroom research projects
 There are no laboratory manuals for OPNET Modeler
Copyright © 2005 OPNET Technologies, Inc. Confidential, not for distribution to third parties.
16
THE END
Copyright © 2005 OPNET Technologies, Inc. Confidential, not for distribution to third parties.