Network Load Generator

1
Network Load Generator

• Michael Gustus
• Dmitriy Mosheyev
• Supervisor Alina Maor

2
Introduction
3
???

• ?????? ?????? ????? ????? ???? ???? ???? "???
  ???? ???? ????? ??????? ??? ??? ?????? ???
  ??????".
• ???? ????? ???? ???? ?????? ?????
• ??????? ???? ??? ?????. ??? ???????? ?? ??? ????
  ?? ???? ??????, ????? ????? ?????? ??? ?????
  ????? ????? ????.
• ??????? ???? ?????? ???? ???? ???? ??? ??????
  ?????????.

4
???

• ?? ???? ????? ?? ?????? ???? ????? ???,
  ?????? ?? ???? ????? ????.
• ????? ??"? ????? ?????? ?? ????? ???? ??????
  ????? ?????? ??????? ?? ????.
• ???, ???? ??? ????? ???? ????? ??? ????? ??? ??
  ??? ????? ????.

5
Project Definition
6
????? ???????

• ????? ????? ?????? ???? ?? ?????? ??????? ?? ???
  ?????? ????? ??????, ????? ????? ??????? ??? ???
  ????? ??? ???? ????.
• ?????? ?????? ???? ???? ?????? ????? ???????
  ????? WINDOWS ?-UNIX.

7
???? ????? ???????
8
????????

• ????? ???????? ????? ????? ???? ?? ??? ??????
  ????? ??????? ???????.
• ??? ?? ???? ?????? ?????? ?? ? ?-e.

9
????????

• ????? ???? ????? ?? IP ?? ????, PORT ?? ????,
  ??? ????? ???? ????? ??????.
• ????? ???????? ????? ????? ???? ?? ??? ??????
  ????? ??????? ???????.
• ??? ??????? ???? ???? ?????? ?????? ?? ? ?-e ???
  ???????? ?????

• - f ??? ?????
• ??? ?-? ????????. ????? ???? ????????? ?-? ??????
  ?????? ?? 10 ?????? ???????.
• ?????? ???? ????? ?????? ?? ? ???. ????? ????
  ???? ????? ???? e??? ?????? ?? ?.
• ????? ????? ?? ????/????? ????? ??? ???? ??????
  ???. ??? ?? ????? ???? ????? ?? ?????? ???????
  ?????? ????? ???? ????? ?? ??? ???? ????? ?? ??
  ??????? ???????.
• ??????? ?????? ???? ???? log.txt

10
Implementation

• ????? ????????? UDP
• ????? ????? ???? ?? ?????? ??????? ??????

• ????? ?????? ?????? ??? Java
• ???? ?????? ?? ?????? ????? ?????
• GUI
• ????? ?-Network Programming

11
UDP protocol

• This User Datagram Protocol (UDP) is defined to
  make available a datagram mode of packet-switched
  computer communication in the environment of an
  interconnected set of computer networks. This
  protocol assumes that the Internet Protocol (IP)
  is used as the underlying protocol.
• This protocol provides a procedure for
  application programs to send messages to other
  programs with a minimum of protocol mechanism.
• The protocol is transaction oriented, and
  delivery and duplicate protection are not
  guaranteed.
• Applications requiring ordered reliable delivery
  of streams of data should use the Transmission
  Control Protocol (TCP).

12
UDP protocol (cont.)

• User Interface
• A user interface should allow
• the creation of new receive ports,
• receive operations on the receive ports that
  return the data octets and an indication of
  source port and source address,
• and an operation that allows a datagram to be
  sent, specifying the data, source and destination
  ports and addresses to be sent.
• IP Interface
• The UDP module must be able to determine the
  source and destination internet addresses and the
  protocol field from the internet header. One
  possible UDP/IP interface would return the whole
  internet datagram including all of the internet
  header in response to a receive operation. Such
  an interface would also allow the UDP to pass a
  full internet datagram complete with header to
  the IP to send. The IP would verify certain
  fields for consistency and compute the internet
  header checksum.

13
Project Structure
14
Main Modules

• Server module
• This module establishes the connection to client,
  computes the delta, generates and sends required
  packets with required interval to client.
• Client module
• This module responsible for initializing the
  connection with server, sending him the desirable
  parameters, receiving the packets from server
• Transfers the packet data to statistics module.
• Generates log file.

15
Main Modules

• Server GUI module
• Displays connection information.
• Client GUI module
• Enables user friendly interface to
• Run simulation.
• View statistics result of specific simulation.
• Export the results to the file.
• Import the results from the file.
• Displays packet information table during the
  simulation.

16
Main Modules

• Statistics module
• Responsible for computing various statistics
  about network load from the log file.
• Number of packet received by the client
• Number of packet lost
• Min packet delta
• Max packet delta
• Average packet delta
• Median packet delta
• Delta average time distribution
• Lost packets time distribution
• Delta average diagram

17
Main Modules

• Statistics GUI module
• Responsible for displaying the statistics
• File controller module
• Responsible for import/export operations.

18
Class Diagram
19
User Guide
20
User Guide

• Pick 2 computers in the network. Start the server
  on one of them, and the client on the another.
• The server shows its IP address and the desired
  port he listens to.

21
Client Window

• In the client window enter the servers IP
  address and port (as seen on the servers
  window).

22
Client Window

• Insert the packet size (in bytes), and packet
  arrival frequency (in byte/sec).

23
Client Window

• To start the simulation, click the Start button.

24
Server Screen
25
Client Screen
The simulation starts running and packet
information is displayed in the table.
26
Statistics Information

• To view the statistics press Show Statistics
  button in the Client Connection Information
  window.

27
Statistics Information

• As a result a Statistics window will be
  displayed

28
Results
29
The issues that were discovered during the
testing

• Javas timer is not sensitive enough
• The minimal time unit is 1 millisecond.
• The largest frequency that can be achieved with
  its timer is one event every 20 milliseconds.
• Thus the biggest frequency we can achieve is 50
  packets per second.
• And since the largest packet size in Ethernet is
  1500 byte therefore we can achieve the maximal
  throughput of 75kb per second.

30
The issues that were discovered during the
testing

• The timer bug in Java with Linux
• When invocating timer on Linux platform with some
  step x the timer invocated with step (x5) msec.
  
• So, than running server on Linux station we
  always get delta bigger than expected.

31
Effect of Frequency Change

• In this test we set the packet size to 500 bytes
  and performed the simulation with different
  frequencies.
• The server was on Linux station and the client on
  the MS-Windows station.

32
Effect of Frequency Change

• On this graph can be seen the influence of the
  frequency on the average delta. While frequency
  rises the delta decreases due to the fact that
  packets are being received more frequently.
• Starting 24000 byte/sec frequency the delta
  converges to some minimal value (this value is
  20ms, the minimal timer step).

33
Effect of Frequency Change
34
Effect of Frequency Change

• Starting the 24000 byte/sec frequency the
  amount of packets that were lost is increases.
• This may happen due to 2 reasons
• The network is overloaded and the packets are
  lost in transmission.
• The frequency is bigger than the speed of CPU
  packet processing time. For example if while CPU
  processed 1 packet were arrived 3 , then at some
  point the buffer will be overloaded and packets
  will start being dropped.

35
One-Many-One Simulation

• On one Windows station we put two Clients that
  were connected to 2 different servers on
  different Linux stations.

• Simulation progress
• Main server started to send packets of 500 bytes
  size with 24000 byte/sec frequency.
• After 2 minutes of work, a second server started
  to send packets of 5000 bytes with 200000
  byte/sec frequency.
• After another 2 minutes the second connection was
  closed and the client continued to receive the
  packets only from the first server.
• On the last minute of simulation we started
  additional connection with the windows client for
  a small time period.

36
One-Many-One Simulation

• First (main) connection with Linux server 1
• Second connection with Linux server 2
• Third connection with Linux server 3

37
One-Many-One Simulation
38
One-Many-One Simulation

• Graphs explanation
• At the moment we started additional client
  connection with another server, the amount of
  packets that were lost has been substantially
  increased (due to the network and clients CPU
  overloading).
• For the same reasons at those times delta also
  has been increased.

39
Traffic disturbance by FTP

• In this test we established a connection from
  Linux server to MS-Windows client with 500 byte
  packets and 24000byte/sec frequency.
• After a while we started to download a file via
  FTP. Average download speed was 120Kb/sec.
• After the download competed we continued the
  session for some time.

40
Traffic disturbance by FTP
41
(No Transcript)