[JDEV] Keep-Alives

Matthias Wimmer m at tthias.net
Thu Jul 4 15:14:30 CDT 2002


Hi Tijl!

Tijl Houtbeckers wrote:

>>TCP keep-alives aren't and won't be needed by the Jabber protocol. It 
>>just solves some problems we have with TCP. Therefore we only need it 
>>with TCP.
>>    
>>
>Well, it's a problem in manny connection oriented enviroment, not just TCP/IP.. if we 
>start using keepalive will people say: we've got the solution now, so why bother on 
>making something on the jabber protocol level (or support it)? 
>
What sort of connection do you think of? AFAIK we don't have other 
connections then TCP/IP and TCP/IPv6 for Jabber at the moment.

>I suppose my worries were a bit overstated, keepalives sound like a good solution 
>(should they be opt-in or opt-out?). I hope they won't screw up my GPRS links but I 
>don't think so.. I doubt there that manny implementations of jabber over something 
>else as TCP/IP anyway, we'll just have to cross that bridge when we get there..
>
I would vote for opt-out because they don't break any client and many 
clients can profit by it. It will work with TCP/IP over GPRS the same 
way as with TCP/IP over Ethernet, PPP, HDLC, Bluetooth or any other 
protocol beneath IP or IPv6.

If have enabled keep-alives at my server (amessage) ... just two lines 
of code that have to be added. I will now lower the keep-alive intervals 
bit by bit. I havn't had any problems yet.
There are three "screws" you can turn: The time of inactivity after that 
the keep-alive is send, the interval for resending keep-alives that 
havn't been acknowledged and the number of unsuccessful keep-alives that 
are treated as a broken connection.


Tot kijk
    Matthias

-- 
Fon: +49-700 77007770		http://matthias-wimmer.de/
Fax: +49-89 312 88654		jabber://mawis@charente.de





More information about the JDev mailing list