Order Tray | Contact Us | Home | SIG Lists

[aprssig] Periodic Disconnects from APRS-IS

AE5PL Lists HamLists at ametx.com
Fri Jan 19 11:51:33 UTC 2007


Scott,

This thread you created is dead.  RFC896 defines the Nagle algorithm,
but doesn't show (as it shouldn't) the net effects.  I have told you
what the net effects are and what the studies done by numerous other
network experts have shown but you refuse to accept this.  So, simply
put: APRS-IS servers and clients should and do disable the Nagle
algorithm and should not buffer packets.  This is a non-issue and simply
a distracter from the content and intent of my original post.  You can
continue your posts on this subject but as far as I am concerned the
subject is dead because it is not going to change.

My original post, which you have successfully distracted me from, bears
repeating: if your client is unable to keep up with the full stream,
whether because of the high amount of processing it has to do (most
common) or because of the limited bandwidth you have, please consider
switching to a server filter port (most servers support port 14580 as a
user-defined filter port).  You will find your connections will be more
reliable and your client more responsive with the reduced packet
processing your client has to do.

73,

Pete Loveall AE5PL
pete at ae5pl.net



> -----Original Message-----
> From: Scott Miller
> Posted At: Thursday, January 18, 2007 10:50 PM
> Subject: RE: [aprssig] Periodic Disconnects from APRS-IS
> 
> Nagle's algorithm alone is not the problem here.  Read RFC896 - done
> right,
> it'll result in only one MSS worth of data being buffered.  Last I
> checked,
> that works out to about 1/10 of a second worth of APRS IS data.  The
> real
> problem is when it's combined with delayed ACKs.  See below.




More information about the aprssig mailing list