CU-DU FLOW CONTROL OPTIMIZATIONS FOR LOSSY MIDHAUL IN O-RAN NETWORKS
A method for optimizing Control Unit (CU)-Distributed Unit (DU) flow control to detect congestion over midhaul and take corrective action to reduce packet dropping in the midhaul includes: computing, by DU, desired data rate (DDR) for each data radio bearer (DRB); and sending, by the DU, the DDR every time when desired buffer size (DBS) is sent from DU to CU user plane (CU-UP), along with Downlink Data Delivery Status (DDDS). DU computes DDR for a DRB as follows: DDR=TBS(MCS″, prb″, rank)*(number of slots/subframe in 1 sec)*(beta)*(beta_drb), with TBS being the Transport block size computation, and MCS″=min {(current MCS+MCS_step), MCSmax}.
1 . A method for optimizing midhaul flow control between a control unit (CU) and a distributed unit (DU) of an Open Radio Access Network (O-RAN), comprising:
computing, by the DU, desired data rate (DDR) for each data radio bearer (DRB); and
sending, by the DU, the DDR every time when desired buffer size (DBS) is sent from DU to CU-user plane (CU-UP), along with Downlink Data Delivery Status (DDDS);
wherein the DU computes DDR for a DRB as follows:
DDR=TBS (MCS″, prb″, rank)*(number of slots in 1 sec)*(beta)*(beta_drb), with TBS being the transport block size computation, and wherein one of:
prb
’
’
=
max
(
PRB
max
#UE
s
,
PRB
avg
,
1
)
wherein current MCS is an instantaneous current value of MCS at a current time instant t, avgMCS current is current weighted average MCS at a current time instant t, MCSmax is the maximum possible MCS, and MCS_Step is a correction factor for the current MCS and the average MCS current , and MCS_Step is one of positive or negative depending on channel quality indicator (CQI) consistency.
2 . The method of claim 1 , wherein for b) MCS″=min {(average MCS(current)+MCS_step), MCSmax}, current weighted average MCS at a current time instant t, avgMCS current , of a user equipment (UE) is computed as follows:
a
)
MCS
’
’
=
min
{
(
current
modulation
and
coding
scheme
(
current
MCS
)
+
MCS_step
)
,
MCS
max
}
,
or
b
)
MCS
’
’
=
min
{
(
average
MCS
current
+
MCS_step
)
,
MCS
max
}
,
wherein alpha is between 0 and 1, and avgMCS previous is a weighted average of MCS at previous time instant t-1.
3 . The method of claim 1 , wherein:
avgMCS
current
=
avgMCS
previous
*
(
1
-
alpha
)
+
current
MCS
*
alpha
,
PRB_avg is the average number of physical resource blocks (PRBs) allocated to UE counting all downlink (DL) slots, #UEs is the number of active UEs in the cell, and PRB max is the maximum number of possible PRBs in the cell.
4 . The method of claim 1 , wherein:
prb
’
’
=
max
(
PRB
max
#UE
s
,
PRB
avg
,
1
)
,
PRB_avg is the average number of physical resource blocks (PRBs) allocated to UE counting all downlink (DL) slots, #UEs is the number of active UEs in the cell, and PRB max is the maximum number of possible PRBs in the cell.
5 . The method of claim 1 , further comprising:
reducing, by the DU, the computed DDR for a DRB by a reduction value δ which depends on an amount of loss of downlink (DL) new radio user-plane sequence numbers (NR-UP SNs) in a specified time interval over a midhaul connection between the CU and the DU.
6 . The method of claim 5 , further comprising:
at least one of:
i) sending, by the DU to the CU-UP, a flag along with the DDDS to indicate the computed DDR has been reduced to reflect the loss of DL NR-UP SNs;
ii) sending, by the DU to the CU-UP, the DDDS after the DU discovers that the amount of loss of DL NR-UP SNs in the specified time interval over the midhaul connection between the CU and DU is above a specified threshold; and
iii) adjusting, by the DU, a frequency of sending the DDDS, depending on whether packet loss ratio estimate at DU based on the amount of loss of DL NR-UP SNs in the specified time interval over the midhaul connection between the CU and DU is above a specified threshold.
7 . The method of claim 1 , further comprising:
reducing, by the CU-UP, the received DDR for a DRB by a reduction value ε which depends on an amount of loss of downlink (DL) new radio user-plane sequence numbers (NR-UP SNs) in a specified time interval over a midhaul connection between the CU and the DU.
8 . The method of claim 1 , further comprising:
computing, by the CU-UP, an effective DDR for a DRB based on the received DDR and information derived from the DDDS regarding a proportion of lost DL NR-UP SNs relative to transmitted NR-UP SNs in a specified time interval over a midhaul connection between the CU and the DU, wherein the effective DDR is the minimum value between i) the received DDR and ii) the proportion of lost DL NR-UP SNs relative to transmitted NR-UP SNs in the specified time interval.
9 . A system for optimizing midhaul flow control in an Open Radio Access Network (O-RAN), comprising:
a centralized unit (CU) comprising a CU user plane (CU-UP); and
a distributed unit (DU) configured to i) compute a desired data rate (DDR) for each data radio bearer (DRB), and ii) send the DDR every time when desired buffer size (DBS) is sent from the DU to the CU-UP, along with Downlink Data Delivery Status (DDDS);
wherein the DU is configured to compute the DDR for a DRB as follows:
DDR=TBS (MCS″, prb″, rank)*(number of slots in 1 sec)*(beta)*(beta_drb), with TBS being the transport block size computation, and wherein one of:
a
)
MCS
’
’
=
min
{
(
current
modulation
and
coding
scheme
(
current
MCS
)
+
MCS_step
)
,
MCS
max
}
,
or
b
)
MCS
’
’
=
min
{
(
average
MCS
current
+
MCS_step
)
,
MCS
max
}
,
wherein current MCS is an instantaneous current value of MCS at a current time instant t, avgMCS current is current weighted average MCS at a current time instant t, MCSmax is the maximum possible MCS, and MCS_Step is a correction factor for the current MCS and the average MCS current , and MCS_Step is one of positive or negative depending on channel quality indicator (CQI) consistency.
10 . The system of claim 9 , wherein for b) MCS″=min {(average MCS(current)+MCS_step), MCSmax}, current weighted average MCS at a current time instant t, avgMCS current , of a user equipment (UE) is computed as follows:
avgMCS
current
=
avgMCS
previous
*
(
1
-
alpha
)
+
current
MCS
*
alpha
,
wherein alpha is between 0 and 1, and avgMCS previous is a weighted average of MCS at previous time instant t-1.
11 . The system of claim 9 , wherein:
prb
’
’
=
max
(
PRB
max
#UE
s
,
PRB
avg
,
1
)
,
PRB_avg is the average number of physical resource blocks (PRBs) allocated to UE counting all downlink (DL) slots, #UEs is the number of active UEs in the cell, and PRB max is the maximum number of possible PRBs in the cell.
12 . The system of claim 9 , wherein:
prb
’
’
=
max
(
PRB
max
#UE
s
,
PRB
avg
,
1
)
,
PRB_avg is the average number of physical resource blocks (PRBs) allocated to UE counting all downlink (DL) slots, #UEs is the number of active UEs in the cell, and PRB max is the maximum number of possible PRBs in the cell.
13 . The system of claim 9 , wherein:
the DU is configured to reduce the computed DDR for a DRB by a reduction value δ which depends on an amount of loss of downlink (DL) new radio user-plane sequence numbers (NR-UP SNs) in a specified time interval over a midhaul connection between the CU and the DU.
14 . The system of claim 13 , wherein:
the DU is configured to send to the CU-UP a flag along with the DDDS to indicate the computed DDR has been reduced to reflect the loss of DL NR-UP SNs.
15 . The system of claim 9 , wherein:
the CU-UP is configured to reduce the received DDR for a DRB by a reduction value ε which depends on an amount of loss of downlink (DL) new radio user-plane sequence numbers (NR-UP SNs) in a specified time interval over a midhaul connection between the CU and the DU.
16 . The system of claim 9 , wherein:
the CU-UP is configured to compute an effective DDR for a DRB based on the received DDR and information derived from the DDDS regarding a proportion of lost DL NR-UP SNs relative to transmitted NR-UP SNs in a specified time interval over a midhaul connection between the CU and the DU, wherein the effective DDR is the minimum value between i) the received DDR and ii) the proportion of lost DL NR-UP SNs relative to transmitted NR-UP SNs in the specified time interval.