RAN driven L4S marking and processing for congestion management in an O-RAN based network architecture
A system and method for reducing bottleneck queues in the Open Radio Access Network (O-RAN) architecture through classification of Packet Data Convergence Protocol (PDCP) and Radio Link Control (RLC) queues at the Centralized Unit-User Plane (CU-UP). The system analyzes and determines a category for the CU-UP PDCP queue for the Data Radio Bearer (DRB), x, at the CU-UP and analyzes and determines a category for the DU RLC queue for the DRB, x, at the CU-UP. If the Congestion Indicator indicates Congestion Experienced (CE) for this DRB in the RAN, the system at the CU-UP starts inserting CE in the IP packet header for packets going in the reverse direction towards the application source, which in turn will modify the rate of transfer based on the number and duration of CE received.
1 . A system for reducing bottleneck queues in the Open Radio Access Network (O-RAN) comprising:
a Centralized Unit-User Plane (CU-UP) server;
an application source coupled to said CU-UP server;
a Distributed Server (DU) coupled to said CU-UP server;
a UE coupled to said DU server;
software executing on said CU-UP comprising a Low queuing Latency, Low Loss, and Scalable throughput (L4S) module that measures occupancy level and average delay in a Packet Data Convergence Protocol (PDCP) queue of said CU-UP server and analyzes a Desired Buffer Size (DBS) and a Desired Data Rate (DDR) received from a Radio Link Control (RLC) queue of said DU server for Downlink (DL) L4S data traffic;
software executing on said DU server that computes DBS and DDR for each L4S Data Radio Bearer (DRB) for DL traffic at the DU using a) radio condition of the UE, b) packet error rate of the UE and the DRB, c) Quality of Service (QoS) requirements of the DRB, and d) cell load and amount of buffer available for the DRB in the RLC queue at the DU server;
the software executing on said DU server communicates DBS and DDR for each L4S DRB to said CU-UP server;
said L4S module at said CU-UP server marks DL packets of a DRB x in the PDCP queue as Congestion Experienced (CE) in a packet header based on occupancy level and average delay of packets in the PDCP queue, and DBS and DDR received from the DU for DRB x;
said L4S module at CU-UP server identifies Uplink (UL) packets of the DRB x sent towards the source of the DL traffic for DRB x and marks the packet header of the identified UL packets as CE, if the L4S module at CU-UP server has marked DL packets for DRB x as CE;
wherein said application source, upon receiving UL packets with CE indications from said Cu-UP server, modifies a rate of data transfer to the CU-UP server based on a number or a duration of CE indications received using scalable congestion control mechanisms;
wherein said L4S module continues indicating CE until:
CE has been indicated in a number of consecutive uplink packets until a number threshold is reached;
CE has been continuously indicated for a time interval until a time threshold is reached; or
the occupancy measurement of the PDCP queue drops below a threshold level or the communicated DBS value from said DU server rises above a threshold level, or both, which results in the L4S module indicating NC for DRB x.
2 . The system according to claim 1 ,
wherein said CU-UP monitors UL and DL GTP-U tunnels over a Next Generation User Plane (NG-U) interface for the DRB x and a corresponding Protocol Data Unit (PDU) session respectively, and
wherein said L4S module identifies UL packets of the DRB x being sent to the source of the DL traffic for DRB x and marks these as CE if the L4S module at CU-UP server has marked DL packets for DRB x as CE,
wherein said L4S module marks DL packets and UL packets in equivalent proportions in a given time interval.
3 . The system according to claim 2 , wherein when said application source receives a CE indication, said application source reduces a data rate at which it sends DL, data towards the UE.
4 . The system according to claim 3 , wherein the data rate reduction is based, at least in part, on a number of CE indications received in uplink packets.
5 . The system according to claim 1 , further comprising:
a Radio Unit near-real-time Radio Intelligent Controller (near-RT-RIC) coupled to said CU-UP server;
said near-RT-RIC receives the classification for the DRB x and analyzes the classification in conjunction with operator policies to determine whether to indicate CE for each DRB carrying L4S traffic.
6 . The system according to claim 5 , wherein if said near-RT-RIC indicates CE, the CE determination is transmitted to said CU-UP server.
7 . The system according to claim 6 , wherein when said CU-UP server receives a CE determination from said near-RT-RIC, said CU-UP server inserts the CE indications in the packet header for packets transmitted to said application source.
8 . A method for reducing bottleneck queues in the Open Radio Access Network (O-RAN) having a Centralized Unit-User Plane (CU-UP) server coupled to an application source, a Distributed Server (DU) coupled to said CU-UP server, a UE coupled to said DU server and software executing on the CU-UP server comprising a Low queuing Latency, Low Loss, and Scalable throughput (L4S) module, the method comprising the steps of:
the L4S module measuring an occupancy level and average delay in a Packet Data Convergence Protocol (PDCP) queue of said CU-UP server;
the L4S module analyzing a Desired Buffer Size (DBS) and a Desired Data Rate (DDR) received from a Radio Link Control (RLC) queue of the DU server for Downlink (DL) L4S data traffic;
the DU server computing DBS and DDR for each L4S Data Radio Bearer (DRB) for DL traffic at the DU using a) radio condition of the UE, b) packet error rate of the UE and the DRB, c) Quality of Service (QoS) requirements of the DRB, and d) cell load and amount of buffer available for that DRB in the RLC queue at the DU server;
the DU server transmitting DBS and DDR for each L4S DRB to the CU-UP server,
the L4S module marking DL packets of a DRB x in the PDCP queue as Congestion Experienced (CE) in a packet header based on occupancy level and average delay of packets in the PDCP queue, and DBS and DDR received from the DU for DRB x;
the L4S module identifying Uplink (UL) packets of the DRB x sent towards the source of the DL traffic for DRB x and marks the packet header of the identified UL packets as CE, if the L4S module at CU-UP server has marked DL packets for DRB x as CE;
the application source, upon receiving UL packets with CE indications from the CU-UP server, modifying a rate of data transfer to the CU-UP server based on a number or a duration of CE indications received using scalable congestion control mechanisms;
wherein the L4S module continues indicating CE until:
CE has been indicated in a number of consecutive UL packets until a number threshold is reached;
CE has been continuously indicated for a time interval until a time threshold is reached; or
the occupancy measurement of the PDCP queue drops below a threshold level or the communicated DBS value from the DU server rises above a threshold level, or both, which results in the L4S module indicating NC for DRB x.
9 . The method according to claim 8 , further comprising the step of:
the CU-UP server monitors uplink and downlink GTP-U tunnels over a Next Generation User Plane (NG-U) interface for the DRB x and a corresponding Protocol Data Unit (PDU) session respectively,
the L4S module identifies UL packets of the DRB x being sent to the source of the DL traffic for DRB x and marks these as CB if the L4S module at the CU-UP server has marked DL packets for DRB x as CE, and
the L4S module marks DL packets and UL packets in equivalent proportions in a given time interval.
10 . The method according to claim 9 , wherein when the application source receives CE indications, the method further comprising the step of:
the application source reduces a data rate at which it sends DL data towards the UE.
11 . The method according to claim 10 , wherein the data rate reduction is based at least in part on the number of indications received in uplink packets.
12 . The method according to claim 8 , wherein a Radio Unit near-real-time Radio Intelligent Controller (near-RT-RIC) is coupled to the CU-UP server, the method further comprising the step of:
the near-RT-RIC receiving the classification for the DRB x and analyzing the classification in conjunction with operator policies to determine whether to indicate CE for each DRB carrying L4S traffic.
13 . The method according to claim 12 , further comprising the step of:
when the near-RT-RIC indicates CE, the CE determination from said near-RT-RIC is transmitted to the CU-UP server.
14 . The method according to claim 13 , further comprising the step of:
when the CU-UP server receives the CE determination, the CU-UP server then inserts the CE indications in the packet header for packets transmitted to the application source.