[Standards-JIG] Re: WHACK

Tijl Houtbeckers thoutbeckers at splendo.com
Wed Apr 26 18:20:02 UTC 2006


On Wed, 26 Apr 2006 19:59:03 +0200, Maciek Niedzielski  
<machekku at uaznia.net> wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Michal vorner Vaner wrote:
>> Well, with whings, you detect the death to drop it by server and keep it
>> forever. However, the timeout on the tcp level is really big (tested on
>> a real, quite used server, it was 15 minutes).
>
> I remember that.. hmm.. 2 years ago I left my session on jabber.org on
> Saturday, and when I returned from holidays on Wednesday, it was still
> active...

On the TCP level there is no such thing a time out for an established  
socket. You can set up a TCP socket and not send a single TCP packet for 2  
years, and start using it again if you want. This also means a  
disconnection can go undetected for 2 years, (as long as you're not  
sending any data). The whole server on the other end could have gone up in  
flames. This is why "whings" (whitespace pings) are currently used (as an  
alternative to keep alive packets, which have the same problems as TCP  
level acks, they're not always accesible and do no necisarly function on  
what from the jabber perspective is hop-to-top).

This is a different problem from what acking tries to solve. "Whinging" is  
for making sure you're not missing anything you could be getting,  
"whacking" is for making sure that what you think you sent was actually  
sent. How to seperate the two is another question.. though I guess using  
XML instead might still be the best answer.




More information about the Standards mailing list