Any counter in Huawei LTE to show E2E packet delay, as new firewall cutover made DL user throughput and traffic volume degradation in our network, anyone have an idea, pls share
PDCP_DU_Delay/LOSS or discard
Yes, packet delay is increased, why it is increased? This is the question, i check 01 another thing PDCP index 8 samples are increased with decreased PDCP index 7 same time of the cutover, but nothing modified at radio end.
Anyone support to check, Index 8 increased and Index 7 decreased after Old to new Firewall cutover, nothing modified at radio end (e-NodeB), how this changed happened? and how to rectify this issue? as Traffic and DL user THB decreased with increased in Packet delay
this is due to change in transmission link.
Index 7: [1216, 1439] bytes
Index 8: [1440, 1503] bytes
Before that, transmission link can not send packet bigger than 1439 bytes, later it can.
I think that bigger packet size will give higher throughput.
Hainm, is any modification needed at e-NodeB end to schedule this bigger packet more quickly?
what is formula of throughput, how is ratio of last TTI traffic?
What is formula of packet delay, if radio packet delay ~ 500ms, I think it’s too large, I suspect of these value.
and This is not problem of enb, no modification is needed at enb.
DL User THB=(L.Thrp.bits.DL-L.Thrp.bits.DL.LastTTI)/L.Thrp.Time.DL.RmvLastTTI/1000
DL Packet Delay=L.Traffic.DL.PktDelay.Time/L.Traffic.DL.PktDelay.Num
It seems last TTI traffic ratio decreased a little. Please collect also in 1 table daily
please check counters about: FEGE DL no of bytes received, packet received & calculate average byte/packet
Before we had Juniper Firewall and we swapped it to Huawei, Huawei is saying before Packet size was 1350 byte in Juniper Firewall and now Huawei default Firewall packet size is 1460 byte which caused scheduling delay at e-NodeB level, that’s why High delay and low DL user Throughput observed at radio level, Is it seems logical??
No, I don’t think it is reasonable. Maybe should be opposite, the larger packet shall have a higher throughput and lower end to end delay, if the priority is maintained.
I think so. I met one case TCP packet size lower then throughput reduced alot.
But here the case is opposite higher MTU packet size reduced the throughput, also we have modified the packet size in Huawei Firewall from 1460 byte to 1350 byte like old Firewall setting but nothing changed.
Id like to add that in many networks we dont even know about the “firewall cutover” event… hehehe! we are never informed. but we are pressured to find the cause… that always is the fact heheh!