Facts to know:
- You use Dos protection by creating Dos policy (Policy & Objects -> IPv4/Ipv6 DoS Policy) in which you enable/modify anomalies.
- The list of anomalies is pre-set in any policy you create. You only have the choice which ones to enable and which ones not to.
- All anomalies are set by default to Pass the offending traffic and are disabled, so make sure under the given anomaly to set
status enableand action to
- You can (actually must) specify: source/destination IPs to match the DoS policy (
allcan be used), service (
ALLcan be used), and incoming interface to apply the DoS policy to.
- Thresholds for anomalies are configurable and do what they say - once traffic matched by this policy exceeds the threshold, it gets blocked. No learning or adaptive thresholds here.
- By default, only exceeding the threshold packets get blocked. To block the sender IP completely, you can use
set qurantineparameter under the specific anomaly.
- Dos sensor/policy protects against INCOMING traffic for the specified interface.
- For smarter anti-DDoS solution Fortinet have FortiDDoS physical appliance.
- Fortigate applies Dos protection early in the policy matching, before the Security policy is checked, so it consumes less resources than blocking the same traffic in Security rules. This means, though, that even if some security rule allows traffic, if such traffic exceeds DoS thresholds it may be blocked.
- Note: in previous versions of FortiOS the feature was called DoS sensor, so I mention it for easier reference only. In FortiOS 6.x and newer it is called DoS Policy.
- From my personal experience, to protect large networks with this DoS feature of Fortigate is more hassle than help. The false positives, especially for TCP SYN and alike protections, would block legitimate clients to the internal servers available from the Internet due to sudden surge of the client requests. You would need then to fix the thresholds, then again... For small networks, and those that do not have accessible from the outside servers, it may be a nice to have feature.
Configuring DoS policy.
I enable just
icmp_flood anomaly here and change the threshold to 10 packets per second sent to destination of 126.96.36.199 :
config firewall DoS-policy edit 1 set interface "port1" set srcaddr "all" set dstaddr "188.8.131.52" set service "ALL_ICMP" config anomaly edit "tcp_syn_flood" set threshold 2000 next edit "tcp_port_scan" set threshold 1000 next edit "tcp_src_session" set threshold 5000 next edit "tcp_dst_session" set threshold 5000 next edit "udp_flood" set threshold 2000 next edit "udp_scan" set threshold 2000 next edit "udp_src_session" set threshold 5000 next edit "udp_dst_session" set threshold 5000 next edit "icmp_flood" set status enable set action block set threshold 10 next edit "icmp_sweep" set threshold 100 next edit "icmp_src_session" set threshold 300 next edit "icmp_dst_session" set threshold 1000 next edit "ip_src_session" set threshold 5000 next edit "ip_dst_session" set threshold 5000 next edit "sctp_flood" set threshold 2000 next edit "sctp_scan" set threshold 1000 next edit "sctp_src_session" set threshold 5000 next edit "sctp_dst_session" set threshold 5000 next end next end
Sending 5 packets per second, traffic is NOT blocked:
root@ubuntu:~# ping -i 0.2 184.108.40.206 PING 220.127.116.11 (18.104.22.168) 56(84) bytes of data. 64 bytes from 22.214.171.124: icmp_seq=1 ttl=255 time=3.03 ms 64 bytes from 126.96.36.199: icmp_seq=2 ttl=255 time=1.96 ms 64 bytes from 188.8.131.52: icmp_seq=3 ttl=255 time=0.469 ms 64 bytes from 184.108.40.206: icmp_seq=4 ttl=255 time=0.318 ms 64 bytes from 220.127.116.11: icmp_seq=5 ttl=255 time=0.405 ms 64 bytes from 18.104.22.168: icmp_seq=6 ttl=255 time=0.497 ms
Sending roughly 10 packets per second - Fortigate starts to block excessive icmp packets.
root@ubuntu:~# ping -i 0.1 22.214.171.124 PING 126.96.36.199 (188.8.131.52) 56(84) bytes of data. 64 bytes from 184.108.40.206: icmp_seq=1 ttl=255 time=1.33 ms 64 bytes from 220.127.116.11: icmp_seq=2 ttl=255 time=0.712 ms ... --- 18.104.22.168 ping statistics --- 143 packets transmitted, 115 received, 19% packet loss, time 14526ms
To see the active attacks/blocked anomalies (block happens when
freq goes 10 or higher):
diagnose ips anomaly list
FG3-AS1680 # diagnose ips anomaly list list nids meter: id=icmp_flood ip=22.214.171.124 dos_id=1 exp=993 pps=1 freq=14
Next, I add second policy with destination address
all but also with qurantine enabled.
config firewall DoS-policy edit 2 set interface "port1" set srcaddr "all" set dstaddr "all" set service "ALL" config anomaly edit "icmp_flood" set status enable set log enable set quarantine attacker set quarantine-expiry 2m <-- to set to 2 min I entered: 000d00h02m set quarantine-log disable set threshold 10 next
Exceeding the threshold:
root@ubuntu:~# ping -c 2000 -i 0.01 126.96.36.199 PING 188.8.131.52 (184.108.40.206) 56(84) bytes of data. 64 bytes from 220.127.116.11: icmp_seq=1 ttl=254 time=0.741 ms 64 bytes from 18.104.22.168: icmp_seq=2 ttl=254 time=1.82 ms 64 bytes from 22.214.171.124: icmp_seq=3 ttl=254 time=1.89 ms --- 126.96.36.199 ping statistics --- 2000 packets transmitted, 11 received, 99% packet loss, time 24308ms
As you can see, 1st 10 packets were allowed, the 11th packet triggered the following block.
FG3-AS1680 # diagnose ips anomaly list id=icmp_flood ip=188.8.131.52 dos_id=2 exp=998 pps=38 freq=83
Also, because I set qurantine period for 2 minutes, even after stopping the attack traffic, the sending server is blocked from sending ANY packets to the target 184.108.40.206 for the next 2 minutes:
root@ubuntu:~# ping 220.127.116.11 PING 18.104.22.168 (22.214.171.124) 56(84) bytes of data. --- 126.96.36.199 ping statistics --- 11 packets transmitted, 0 received, 100% packet loss, time 10029ms
Releasing the blocked senders
Fortigate does not show us the source IPs of the blocked hosts, just the target IP, still, we can clear the blocked attackers list and allow the blocked senders to pass through. If they again send the excessive traffic, they will be blocked again, i.e the clear action is real-time and not permanent. Also, for the senders blocked with the quarantine, clearing the list will still keep them blocked until the qurantine expiration.
FG3-AS1680 # diagnose ips anomaly list list nids meter: id=icmp_flood ip=188.8.131.52 dos_id=1 exp=999 pps=2 freq=20
Clear the list:
diagnose ips anomaly clear
FG3-AS1680 # diagnose ips anomaly clear FG3-AS1680 # diagnose ips anomaly list list nids meter: total # of nids meters: 0.