TCP Round Trip Time to remote address shows 1ms when in fact it should be around 75ms

[JoseSpoo]

I'm working with a customer that has a Fortinet firewall with a PC behind it for testing. 

When running a TCP application test, I'm getting a round trip time consistency of 1% with the the minimum round trip time of 1ms and maximum of 75ms.  This skew in round trip time is causing TCP to burst above what should normally be possible for the latency between the PC behind the Fortinet and the remote address. 

In the past I've repaired this type of issue with a Fortinet by changing a TCP setting within the CLI, but I do not remember what the command was. 

[DaviWang1]

Sounds like you have some proxy based UTM profiles enabled which makes FortiGate running in proxy mode to respond TCP connection on behalf of the server.

If you only have flow based UTM profiles in the affected firewall policy, or don't enable UTM, the response time should be consistent with what you can get without FGT in the middle.