Re: [ns] Wireless packet size limit + wireless bandwidth question

[Filippos Kolovos]

-

On 7/24/06, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
>
>
> -- Forwarded message --
> From: Heo Heo <[EMAIL PROTECTED]>
> To: ns-users@ISI.EDU
> Date: Sat, 22 Jul 2006 17:18:26 -0700 (PDT)
> Subject: [ns] Wireless packet size limit + wireless bandwidth question
>   Hi all,
>
>   I was doing a simple wireless simulation in which 2 nodes send are
> placed next to each other, with one being the sender and the other being
> the receiver (based on the Simple.tcl on ns website). The sender sends
> 500 packets to the receiver.
>
>   1) I tried to change the packet size from 512 to 1024 bytes. I parsed
> the trace file I find out that:
>   - If the packet size is 512 bytes, 500 packets were sent.
>   - If the packet size is 1024 bytes, 1000 packets were sent.
>
>   It seems to me that some layer has automatically split up the 1024-byte
> packets
> into smaller parts.
>
>   So, my question is: what is the default wireless packet size, and how
> to change it ? and Which layer actually split up the big packets ?


The network layer. The splitting-process depends upon the MTU (Maximum
Transmission Unit) used by the
network layer of the network from which the packet passes. For a very draft
explanation, this is the default maximum packet size that a sender can put
onto the channel for transmission, which has been agreed according to a
maximum. For example, the Internet de facto standard for the MTU is 576
bytes,
while for Ethernet networks the maximum is 1500 bytes (to account for
typical TCP packets). However, this can change per connection,
according to the initial 3-way handshake parameters exchange.

The NS-2 simulator uses an MTU of approximately 576 bytes, so this is
probably the reason of the packet-doubling.

For more information concerning the MTU, check the following site:


http://searchnetworking.techtarget.com/sDefinition/0,,sid7_gci213605,00.html

and

http://searchnetworking.techtarget.com/sDefinition/0,,sid7_gci830944,00.html

  2) My next question is about the wireless bandwidth.
>
> By default, ns uses a wireless link bandwidth of 2 Mbps. By
> changing the packet interval, I have at most 670 Kbps. Because there are
> only
> 2 nodes, the 2 nodes must be using full bandwidth for this. Given that
> there are some overhead (RTS/CTS, packet headers, etc.), I would think
> I should get about 1.5Mbps. So, why is the actual speed so much lower
> than that ? Is it natural ?



Did you account for re-transmission errors and the error rate in general? It
is a very common thing
the error rate factor to be somehow high in wireless networks.


  Thanks a lot for your help.


Hope that I've helped you.

-Fk

  Heo
>
>
>
>
> -
>



-- 
Filippos N Kolovos

Software Systems Analyst & Engineer
M.Sc. (Eng.) in Data Communications

Automation & Networking Department
University of Macedonia Library
Egnatia 156, P.O.Box 1591
540 06 Thessaloniki, Greece

E-Mail: [EMAIL PROTECTED],
   [EMAIL PROTECTED]
--

Re: [ns] Wireless packet size limit + wireless bandwidth question

[Phil Miller]

Heo,

Have a look at
.
It answers your direct questions.

You may also wish to look at

if you're interested in adjustments to simulate more recent chipsets accurately.

Phil

On 7/22/06, Heo Heo <[EMAIL PROTECTED]> wrote:
>
>   Hi all,
>
>   I was doing a simple wireless simulation in which 2 nodes send are
> placed next to each other, with one being the sender and the other being
> the receiver (based on the Simple.tcl on ns website). The sender sends
> 500 packets to the receiver.
>
>   1) I tried to change the packet size from 512 to 1024 bytes. I parsed
> the trace file I find out that:
>   - If the packet size is 512 bytes, 500 packets were sent.
>   - If the packet size is 1024 bytes, 1000 packets were sent.
>
>   It seems to me that some layer has automatically split up the 1024-byte 
> packets
> into smaller parts.
>
>   So, my question is: what is the default wireless packet size, and how
> to change it ? and Which layer actually split up the big packets ?
>
>   2) My next question is about the wireless bandwidth.
>
> By default, ns uses a wireless link bandwidth of 2 Mbps. By
> changing the packet interval, I have at most 670 Kbps. Because there are only
> 2 nodes, the 2 nodes must be using full bandwidth for this. Given that
> there are some overhead (RTS/CTS, packet headers, etc.), I would think
> I should get about 1.5Mbps. So, why is the actual speed so much lower
> than that ? Is it natural ?
>
>   Thanks a lot for your help.
>
>   Heo

[ns] Wireless packet size limit + wireless bandwidth question

[Heo Heo]

  Hi all,
   
  I was doing a simple wireless simulation in which 2 nodes send are 
placed next to each other, with one being the sender and the other being 
the receiver (based on the Simple.tcl on ns website). The sender sends 
500 packets to the receiver.
   
  1) I tried to change the packet size from 512 to 1024 bytes. I parsed 
the trace file I find out that:
  - If the packet size is 512 bytes, 500 packets were sent.
  - If the packet size is 1024 bytes, 1000 packets were sent.
   
  It seems to me that some layer has automatically split up the 1024-byte 
packets 
into smaller parts. 
   
  So, my question is: what is the default wireless packet size, and how 
to change it ? and Which layer actually split up the big packets ?
   
  2) My next question is about the wireless bandwidth. 
   
By default, ns uses a wireless link bandwidth of 2 Mbps. By 
changing the packet interval, I have at most 670 Kbps. Because there are only 
2 nodes, the 2 nodes must be using full bandwidth for this. Given that 
there are some overhead (RTS/CTS, packet headers, etc.), I would think 
I should get about 1.5Mbps. So, why is the actual speed so much lower 
than that ? Is it natural ?
   
  Thanks a lot for your help.
   
  Heo
   



-
Talk is cheap. Use Yahoo! Messenger to make PC-to-Phone calls.  Great rates 
starting at 1¢/min.