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
FYI, I have not tested this in snappi, issue found with tgen, assuming implementation is the same.
If you have a traffic item with 16 flow groups and configure the tgen.flow to run until 10k packets are sent ( duration=Duration(FixedPackets(packets=packet_count) ) the traffic item is actually set to have 'Fixed Packet Count' 625 (10k/16 flow groups). This is fine for the scenario where the user wants a total of 10k packets sent no matter the flows/ports.
But we need to consider the scenario where the user actually wants to know how many packets it will be sending per port (which I believe will be more common). This user would expect that when setting packet_count=10k, he will see 10k per traffic item (and from that he could determine per port distribution). Right now this user would have to do the reverse math: if I want to send 10k per flow group, then I need to configure tgen to send 10K x flow groups...
So to overcome this, instead of having tgen silently doing the math, it should take the packet_count as is, and expose a setting for frame count distribution (and let the user decide how to split this count among flow groups).
The text was updated successfully, but these errors were encountered:
The intent of the models is that a user specifies a packet count per flow. This means if there is one flow and a packet count of 100 packets that is exactly what will be sent out - it is an aggregate.
There is no concept of traffic items, flow groups etc as those are implementation specific.
FYI, I have not tested this in snappi, issue found with tgen, assuming implementation is the same.
If you have a traffic item with 16 flow groups and configure the tgen.flow to run until 10k packets are sent ( duration=Duration(FixedPackets(packets=packet_count) ) the traffic item is actually set to have 'Fixed Packet Count' 625 (10k/16 flow groups). This is fine for the scenario where the user wants a total of 10k packets sent no matter the flows/ports.
But we need to consider the scenario where the user actually wants to know how many packets it will be sending per port (which I believe will be more common). This user would expect that when setting packet_count=10k, he will see 10k per traffic item (and from that he could determine per port distribution). Right now this user would have to do the reverse math: if I want to send 10k per flow group, then I need to configure tgen to send 10K x flow groups...
So to overcome this, instead of having tgen silently doing the math, it should take the packet_count as is, and expose a setting for frame count distribution (and let the user decide how to split this count among flow groups).
The text was updated successfully, but these errors were encountered: