Re: Block SSH brute force

In case of abuse or massive traffic we usually block /24 or /16 nets - first for one hour, in case of /24 nets we block for one day.
Most (95%) is china or USA. Usually this does not hit any "good guy", mostly it's from data centers with rented spam servers,
or from the usual provider network ranges where lots of infected PCs are around.

Manually maintaining such lists is not easy, and trusting external lists also has caveats (see: Spamhaus).

In the end - most automatic blocking is # of connections per source ip / net as well as traffic shaping based on serveral rules and spread traffic on multiple physical connections. Once we routed some SSH attackers through our satellite links. The latency of 700ms made them giving up after 30 mins ;-)

-h

Am 11.02.2016 um 17:24 schrieb DrFranken <midrange@xxxxxxxxxxxx>:

In our case we'd need a bigger inbox!!! We also do this for other things such as Telnet, FTP, etc. We also feed some syslog data into the process. Our blocked IP list is 10s of thousands long. I will say that once we started blocking the hits drop off dramatically because we don't block ports we block addresses.