Transcript "The Distance learning and the networking technologies".

"Distance Learning and
Networking Technologies"
Assistant Prof. Dr.-Stelios Savaidis
Department of Electronics,
TEI Piraeus, Greece
Email: [email protected]
Tel:+30 210 5381181
Distance Learning and Network Performance
 Distance Learning relies on the existing network infrastructure

PSTN/ISDN, LANs/WLANs, MANs/WMANs, WANs.
 Network Performance Issues can severely restrict the efficiency of
Distance Learning Applications

QoS issues e.g. throughput, delay etc.
 Network Resource Requirements vary depending on the type of the
Distance Learning Applications

HTTP (best effort), FTP (Variable Rate), Video Streaming (Real Time
Variable Rate) etc.
Distance Learning and Network Performance
 A global approach of Distance Learning should take into account
Network Performance Issues

Degraded Network Performance results to a low QoS from user’s point of
view e.g. slow response to user’s actions.
 An Accurate Network Performance Analysis Provides Proper Network
Dimensioning Decisions that ensure Distance Learning Efficiency at a
minimum technical and/or economical cost.
 Prediction of Network Performance is a rather complicated task but can
be carried out, even for complicated networks, with Network Simulation
Software Packages
OPNET Network Simulator
 OPNET is a powerful event based Network simulator

Produces numerous random traffic events and collects network performance
measurements.
 OPNET Simulations of both Commercial and Generic Network nodes are
available.

Several industrial telecommunication nodes (routers, switches, modems from Cisco,
3 Com etc) are included in OPNET’s library.

Several generic telecommunication nodes can be configured offering customized
solutions.
 OPNET Simulations of Typical Network Links are available

ISDN, ADSL, E1/T1, Ethernet, SDH etc.
 Different Types of Applications are available and can be configured accordingly

HTTP, FTP, Mail, Voice etc.
 Stochastic Models can be used in order to simulate user’s behavior in terms of
traffic generation

Uniform, Exponential, Poisson etc.
Distance Learning Study Case
 Study Case Input Data
2nd Technical School
 Web based Distance
Learning Environment.
3nd Technical School
1st Technical School
 Users from a complex
of 6 high schools.
The Internet
 HTTP and FTP are the
typically
supported
applications.
Administrator
eLearning Web Server
High School
5th Technical School
4th Technical School
Distance Learning Study Case
 Possible Output Data
2nd Technical School
 Maximum number of users
for a given network infrastructure.
 Possible Network Upgrades for a given network
infrastructure e.g.
 what kind of LAN
 what type of DSL to/from
each school.
 what kind of fractional E1
line to/from the Web
Server.
 What kind of Web Server
3nd Technical School
1st Technical School
The Internet
Administrator
eLearning Web Server
High School
5th Technical School
4th Technical School
OPNET Simulation of the Distance Learning Study Case
 An Ethernet LAN represents each school
 The number of users is
variable
 FTP/HTTP traffic is generated randomly.
 FTP file and HTTP page
attributes are also random
 Different types of ADSL link
can be used i.e. with
different upload/download
rates
 Fractional E1 can be used
to/from Web Server
Profile Configuration
Application Configuration
FTP Configuration
HTTP Configuration
HTTP Page Properties
Web Server Configuration
Web Services
 Access to/from Web
Server over a 2 Mbps E1
line
 FTP and HTTP
Services suported
Access to/from User’s Site
User’s Profile
Network Performance Results: Response Time
(ADSL 384/128)
FTP Response Time (sec)
HTTP Page Response Time (sec)
Reasonable Performance ???
Network Performance Results: Response Time
(ADSL 512/128 Kbps)
FTP Response Time - HTTP Page Response Time
Reasonable Performance ???
Network Performance Results: Response Time
(ADSL 1,024/256 Kbps)
FTP Response Time – HTTP Page Rasponse Time
Reasonable Performance ???
Network Performance Results: Throughput at User’s Site
ADSL 384/128 Kbps Throughput
Network Performance Results: Throughput at User’s Site
ADSL 512/128 Kbps Throughput (bits/sec)
Network Performance Results: Throughput at User’s Site
ADSL 1,024/256 Kbps Throughput (bits/sec)
Network Performance Results: Utilization at User’s Site
ADSL 384/256 Kbps Utilization
Network Performance Results: Utilization at User’s Site
ADSL 512/128 Kbps Utilization
Network Performance Results: Utilization at User’s Site
ADSL 1024/128 Kbps Utilization
Network Performance Results: Web Server Throughput
E1 2,048/2,048 Kbps Throughput
(eLearning Server to IP Cloud Communication Link)
Network Performance Results: Web Server Utilization
E1 2,048/2,048 Kbps Utilization
(eLearning Server to IP Cloud Communication Link)
Network Performance Results: Study Case Overview
 Obviously, ADSL with higher data rates increase the number of
eLearning users.
 QoS measurements (e.g. FTP Response Time and HTTP Response
Time) must be counted in order to evaluate network resources vs
number of users.
 Considering HTTP Response Time as the QoS threshold we may
conclude to the following statements
 384/128 Kbps ADSL line can offers an acceptable QoS for 30 users.
 512/128 Kbps ADSL Line increase the number of users to 40 (in
terms of QoS).
 1024/256 ADSL line can offer acceptable QoS to less then 50 and
more than 40 users.
 Web Server E1 link, seems to be an unreasonable choice. A
fractional E1, which is much more affordable for the schools’ budget,
should be used.
"Distance Learning and
Networking Technologies"
Thank you for your attention !!!