TCP throughput?

[Guus Sliepen]

On Tue, Jan 26, 2010 at 11:45:15PM +0900, Masateru KUWATA wrote:

> I am testing tinc on Windows and found some interesting behavior.
> 
> I used "TCPOnly" parameter for both end because I wanted to go through
> NAT. However, by using TCPOnly parameter, the response of PING slowed
> down significantly.
> 
> The PING response with UDP is about 4ms, but it bumped up to 2000ms with
> TCP.
> 
> I am just curious whether this is caused by using TCP.

Yes, this is caused by TCP. However, if you consistently get very large
(>1000ms) ping times, it is probably because there is a lot of congestion
(maybe there is lots of other traffic on the same network), or a lot of packet
loss (bad reception or cable, or a bad ISP).

With tinc 1.0.10 and later, you should not have to use the TCPOnly option
anymore, since it will now automatically detect whether communication via UDP
is possible, and will fall back to TCP when it isn't.

-- 
Met vriendelijke groet / with kind regards,
     Guus Sliepen <guus at tinc-vpn.org>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://www.tinc-vpn.org/pipermail/tinc/attachments/20100127/2e4fef9a/attachment.pgp>