[Standards-JIG] Re: What happened to the ACK proposal?

Gaston Dombiak gaston at jivesoftware.com
Tue Aug 23 17:53:53 UTC 2005


Hey guys,

Thank you all for your help and time spent on this matter. Last week I was 
away on vacation so I missed the opportunity to reply on time. Anyway, after 
reading all the messages my conclusion is that we need some kind of ACK 
solution. It also seems that it should be defined in an RFC and not in a 
JEP.

For now I'm going to add a heartbeat (with whitespace) to detect sooner that 
the connection dropped dead. This solution won't avoid sending packets to 
"black holes" but at least it will help us detect that there is a black hole 
sooner. :) I'm also not discarding the option to implement a custom ACK 
solution that may be used as a basis for a future ACK solution.

Thanks,

  -- Gato

"Gaston Dombiak" <gaston at jivesoftware.com> wrote in message 
news:ddjhl5$peo$1 at sea.gmane.org...
> Hey guys,
>
> I noticed that under certain circumstances Java is not able to detect that 
> a connection was closed. In particular, this problem happens when a 
> router/firewall/NAT is involved in the connection. This problem is 
> affecting us when using s2s communications but in fact it may happen 
> between server and clients too. The source of the problem is that the TCP 
> layer is not throwing an exception even if the server writes data to the 
> socket or is waiting to read from the socket. Therefore, the Java virtual 
> machine is not able to detect the dead connection until lots of data was 
> written to the socket thus making the TCP layer to finally throw the 
> exception. I'm not sure if this problem affects other programming 
> languages but anyway I was wondering why the ACK JEP proposal was 
> rejected. Or  how are other servers (besides Jive Messenger) dealing with 
> this problem.
>
> Regards,
>
>  -- Gato 






More information about the Standards mailing list