|-----Original Message-----
|From: Simon Perreault [mailto:[email protected]] 
|Sent: Thursday, August 11, 2011 8:38 PM
|To: DENG Xiaohong ESP/PEK
|Cc: [email protected]; [email protected]; [email protected]
|Subject: Re: [Softwires] Clarification of the 
|stateles/stateful discussion
|
|On 2011-08-11 07:01, [email protected] wrote:
|> But I don't understand will the figure be different with 'scattered'
|> port-set NAT? To my knowledge, it's something todo with app's 
|> behaviours and NAT type(EIM in our test). Would you explain 
|if I miss 
|> something? Thanks
|
|As it was discussed in this thread [*], an EIM NAT will, in 

Understood EIM behaviors. My point is port-range
NAT or scattered port-set NAT is nothing to do with EIM.

|theory, only be able to support a number of sessions roughly 
|equal to the number of external ports. That is because 
|internal hosts typically do not reuse source port numbers for 
|sessions to different destinations.

Could you explain how this conclusion was drawed?

According to [a], port usage in figure 3 and session usage in figure 4 show 
that,
for Google Earth, port and session usage are roughly the same,
while firefox session usage is roughly 1.5 times as port usage.

|
|If your tests show otherwise, it would be very important for 
|this working group to understand exactly why theory does not 
|match practice.

BitTorrent and Skype are definitely cases that show othersiwse.

As the detailed analysis I stated in the previous email, but don't mind to
explain again:

When BT Client initiates a downloading it first uses the same 
source port to connect to the different destinations (destination IP and port) 
therefore a source port multiplexing different sessions. Data shows maxium
port usuage reaches 160, while session usuage continously stays above 450.

Skype is another example that uses a source port to multiplex 
different sessions thereby introducing less source port than sessions on NAT.
Data shows less than a couple of ports achieve more than a hundred sessions!

But, don't ask me why, as  niether did I write code for BT Client nor for 
Skype. ;0

[a] http://opensourcev6transtechnologies.weebly.com/experiments-results.html


Cheers,
Xiaohong
---
Open-source PCP Client : http://opensourcev6transtechnologies.weebly.com/
A+P experiments results: 
http://opensourcev6transtechnologies.weebly.com/experiments-results.html
IETF81 PCP demo page: 
http://opensourcev6transtechnologies.weebly.com/ietf-81-pcp-demo-site.html
|
|[*] 
|http://www.ietf.org/mail-archive/web/softwires/current/msg02461.html
|
|Simon
|--
|DTN made easy, lean, and smart --> http://postellation.viagenie.ca
|NAT64/DNS64 open-source        --> http://ecdysis.viagenie.ca
|STUN/TURN server               --> http://numb.viagenie.ca
|
_______________________________________________
Softwires mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/softwires

Reply via email to