Delay in packet pass-through

FastL4 has no ability to delay packets it handles. Is there a significant difference in the network path, and are you measuring RTT or the total time for a transfer?

Did you really mean 1 second (1000 ms) , or 1 millisecond ?

It is 1 second (1000ms) delay. We are calculating the total time for transfer. I have seen tcp out of order packers in the capture. Does it make any difference in F5 to forward to destination. Since we have PVA Acceleration enabled in Fast L4 by default, seems it should not delay theoretically, isn’t it?

There will be a small delay while passing through any piece of networking equipment, but this should be in the order of less than a millisecond.. we certainly don't (and can't, as Amy's pointed out) have any way to delay a packet for 1000 milliseconds in the fastL4 profile. I suggest you carefully compare the network paths taken between packets that go via the LTM, and those that do not. You could also run tcdpump on the LTM and look at the timestamps of the packets arriving on the client side, compared with the timestamp of the same data leaving on the server side. Or capture one hop away from the LTM, then two, etc. See if you can find what's happening that is different from the path that works.

Yes. We are not expecting 1 sec dely while communicating to F5 or any other network equipment. I have analysed the packet data. There are so many TCP re-transmission and out of order packets in the pcap which is captired from BIG IP but those are not present in the source captrure. Strange..