Is there any data on packet duplication?

Hal Murray hgm+nanog at
Tue Jun 23 07:51:56 UTC 2020

bill at said:
>  NTP you say? How does iburst work during initial sync up?

How does it work, or how should it work?  1/2 :)

NTP has been around for a long time.  It looks very simple, so anybody thinks 
they can toss off an implementation without much thought.  It will probably 
work, mostly.

The response from an NTP server includes a timestamp that the client put into 
the request.  The client can use that to reject delayed responses to a 
previous request.

When I first started looking for duplicates, I found lots of them.  They were 
NTP version 1 requests.  NTP is up to version 4.  Version 1 came out in 1988, 
RFC 1059.  Since the requests are identical, there is no way for the client to 
separate expected responses from delayed responses from a previous request.

Does anybody happen to know what equipment or software or OS/distro is sending 
version 1 requests?

These are my opinions.  I hate spam.

More information about the NANOG mailing list