[ntp:hackers] Use MSG_CONFIRM flag to keep OS ARP table up to date?

Hal Murray hmurray at megapathdsl.net
Wed May 2 21:06:53 UTC 2012



heiko.gerstung at meinberg.de said:
> performance and reduce unnecessary ARP requests), but in general I would
> say that ntpd should follow the - IMHO - approach that is recommended by
> the OS. On the arp(7) manpage the MSG_CONFIRM flag is mentioned when it
> comes to UDP traffic. 

I think it's more complicated than that.

I think the MSG_CONFIRM flag would make sense on the client side if you knew 
you had received a response to a previous recent request.  That's not the 
case for ntpd acting as a server.

Using it on the server side introduces a chance to break things in subtle 
ways which will be hard to debug.  So I vote against adding it to the stock 
code.  If enough people are doing stress testing, I'd be happy if it was 
added as a configure option so they don't have to hack the sources to run 
their tests.


The original message said
> running stress tests with more than 100k simulated NTP clients

In the real world, how many of those 100k clients would be on the local LAN 
rather than going through a router?

Is arp traffic a problem in the real world?




-- 
These are my opinions, not necessarily my employer's.  I hate spam.





More information about the hackers mailing list