Skip to main content
AndreasK
New Member
August 16, 2011
Question

Activating DoS Protection - Performance Issues

  • August 16, 2011
  • 4 replies
  • 4697 views
Hello, We are using Fortigate 620B to our company (v.4 MR2) only as a Firewall (UTM features are not enabled). We would like to enable DoS protection for 5 public systems. Is this possible without enabling the UTM features? If we have to enable the UTM features (e.g. the IPS) will this affect the overall performance of the device. Currently the CPU usage is below 5% and memory usage is below 15%. Kind regards,

    4 replies

    ede_pfau
    SuperUser
    SuperUser
    August 16, 2011
    Hi, DoS protection IS a UTM feature and part of IPS (combined with rates). Of course it will affect performance, but what are we talking about here, in terms of WAN throughput? Or do you use the FGT for heavy internal traffic as well? The FG-620B is rated at 16 Gbps firewall throughput and 1 Gbps IPS throughput. That should be sufficient for most WAN lines. Design hints: - be sure to select only relevant traffic in the DoS sensor, i.e. TCP only, HTTP/port 80 - scanning SSL / HTTPS as well will cost more - if the CPU load exceeds 50% you might consider putting the DoS sensor into a DoS interface policy. IPS on an interface takes action very early, before firewalling or AV processing, and thus preserves ressources. But frankly I cannot imagine that a DoS sensor on a WAN line challenges the 620B a bit.
    AndreasK
    AndreasKAuthor
    New Member
    August 17, 2011
    Hello Ede, Thanks for the reply! We have activated a DoS Policy to the traffic coming from the external Interface and everything works OK! The CPU and memory usage haven' t increased at all. The traffic to our organization (Internal and external) is very low so the 620B is more than enough. We are monitoring the attack logs in order to fine tune the thresholds. Kind Regards, Andreas
    ede_pfau
    SuperUser
    SuperUser
    August 17, 2011
    You can try to trigger the DoS sensor by temporarily lowering the threshold (e.g. for ICMP flood) and sending in traffic. The detection alone won' t protect you much. Combine with a period of source IP blocking (say, 10 mins) to effectively suppress attacks. We use it a lot to protect from automated login attempts to ssh servers.
    Degies
    New Member
    August 21, 2011
    I am running V4 mr3 P1 and I know there is a bug with the proxy agent on the platform. this is not the proxy options, on the device, but a handling process in the internal working. You might need to Check with Forti if it the same issue