When a client tries to connect to a domain which is on the blacklist for command-and-control, he first starts a DNS query to our internal DNS Server. The DNS Server then starts his own query (in our environment directly to the DNS Server from the ISP). So this traffic is between the two DNS Servers, but because the domain name and the IP of the command-and-control server are in the traffic the query and the response the UTM generates the alert and lists our internal DNS server as infected host (a few pictures can be found here, sorry it’s in german)
In my opinion, it would be better to generate the alert and stop the connection in the next step, when the client REALLY tries to connect to the server and not during the DNS request to another DNS Server. Then the client would be the infected host in the logs. I had to enable DNS logs on the server to identify the affected host.
I am not sure if my considerations are right. If yes it would be a bug (the dns query without connecting to the command-and-control server shouldn’t generate an alert). Thanks for any help.
This thread was automatically locked due to age.
