[aprssig] KPC3+ induced errors => IS dupe checking

[Heikki Hannikainen]

On Wed, 16 Mar 2011, Jason KG4WSV wrote:

> On Wed, Mar 16, 2011 at 2:28 PM, Larry Cerney <lcerney at viawest.net> wrote:
>> where in the APRS system is dup checking
>> done?
>
> everywhere - igates and APRS-IS servers.

This is actually a bit incorrect - the igates are not supposed to do 
duplicate checking, just the APRS-IS servers. I believe that most 
implementations follow this. Please see the beginning of this spec:

http://www.aprs-is.net/IGating.aspx

I find this a good strategy, it's good that the APRS-IS servers implement 
the dupe checking in an uniform way, instead of all igates doing it on 
their own - it reduces the amount of "almost correct but slightly 
different" filter implementations, since there are so many different igate 
packages out there. It also makes it easier to tune the filtering 
algorithms when needed, since the IS servers can be upgraded more easily 
than all the igates. Also, the IS servers could, at some point, find the 
duplicate packets useful, for message routing or something.

For those of you looking at the aprs.fi raw packets list, and finding red 
"duplicate packet" warnings there: aprs.fi also does check for duplicate 
packets. APRS-IS servers do dupe checking over a range of past 30 
seconds - the duplicates filtered in the APRS-IS will never make it to 
aprs.fi and are not shown in the raw packets listing at all. aprs.fi goes 
a bit further, especially for position packets, and tries hard to catch 
the packets delayed by the KPC-3+'s and others - those are shown with the 
red error messages in the raw packets list.

http://blog.aprs.fi/2008/03/on-duplicate-and-delayed-packets.html

   - Hessu