> I tried employing limited and kod, but this didn't keep the ill
> behaving NTP clients from flooding my servers with requests.  So I
> implemented some rate limiting outside of the application.

Back when I had a host in the pool, I did likewise.  Here are the
number of times it tripped, by month, from 2012-12 (which is as far
back as it's convenient to go at the moment) to present.  I stopped
serving NTP on my (then-)pool address at the end of August and of
course got auto-dropped from being advertised shortly thereafter.

2012-12 378
2013-01 379
2013-02 433
2013-03 361
2013-04 455
2013-05 352
2013-06 487
2013-07 571
2013-08 512
2013-09   0
2013-10   0
2013-11   0
2013-12  14

Actually, those are number of log lines; a single incident sometimes
produces multiple log lines, but I just had a quick look, and it
definitely is many incidents per day; I see 2910 distinct <day,IP>
pairs, so somewhere around 7.5 trips per day.

I find it interesting that the misbehaviour stopped almost immediately
upon my stopping NTP service on that address.  (It would be difficult
for me to tell how promptly NTP traffic stopped.)  I also find it
interesting that the misbehaviour stopped even though that host is
still serving NTP on its other addresses (which are nearby, in the same
/16 as, the pool address); whether the floods are coming from abusive
clients or DDoS attempts or what, they clearly are using the pool to
get their addresses.

/~\ The ASCII                             Mouse
\ / Ribbon Campaign
 X  Against HTML                [email protected]
/ \ Email!           7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B
_______________________________________________
pool mailing list
[email protected]
http://lists.ntp.org/listinfo/pool

Reply via email to