You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We start seeing a different result between the Tofino model and the real hardware.
For example, switch(with fabric-spgw-int profile) sends packets with the same 5-tuple to different next-hops on the hardware switch, but in the PTF test with Tofino Model(FabricIPv4UnicastGroupTest*) pass since a while ago(no longer see the failure since end of 2020, see #98)
It is useful to run PTF tests with real hardware to make sure we got what we want.
The text was updated successfully, but these errors were encountered:
@Yi-Tseng do you know why we are not observing this behavior when running PTF-TV on the hardware switch? Could it be that the specific case you were observing on the hardware switch is not covered by PTF, e.g., the switch is processing packets that have different headers than the one used in PTF, such as SCTP from the eNB to the MME?
Yi-Tseng
changed the title
Run PTF test with real hardware
Run PTF test on real hardware
Apr 27, 2021
We start seeing a different result between the Tofino model and the real hardware.
For example, switch(with fabric-spgw-int profile) sends packets with the same 5-tuple to different next-hops on the hardware switch, but in the PTF test with Tofino Model(FabricIPv4UnicastGroupTest*) pass since a while ago(no longer see the failure since end of 2020, see #98)
It is useful to run PTF tests with real hardware to make sure we got what we want.
The text was updated successfully, but these errors were encountered: