Reliability enhancements for multi-access traffic management
The present disclosure is related to multi-access traffic management in multi-access computing environments. A reliability enhancement engine (REE) operated by a multi-access multipath traffic manager identifies traffic flows that have high reliability requirements (“reliability flows”) and performs admission control for the reliability flows. The REE determines a coding rate (e.g., a redundancy factor) and traffic distribution strategies for the identified reliability flows. The REE applies cross-access network coding (NC) on the reliability flows, and handles multi-access acknowledgements and traffic pacing, which may involve multiplexing traffic of the reliability flows with non-high-reliability traffic. Other embodiments may be described and/or claimed.
1. An edge compute node for providing reliability enhancements in a multi-access communication network, the edge compute node comprising:
interface circuitry arranged to communicatively couple the edge compute node with a set of network access nodes (NANs) in the multi-access communication network, wherein one or more NANs in the set of NANs is communicatively coupled with one or more user devices; and
processor circuitry connected to the interface circuitry, wherein the processor circuitry is to:
collect, via the interface circuitry, metrics related to the multi-access communication network from one or more data sources within the multi-access communication network;
determine, based on the collected metrics, a reliability target, a network coding (NC) group size, and per path characteristics for each network path of a set of network paths in the multi-access communication network;
determine a traffic distribution strategy based on the per path characteristics, the NC group size, and the reliability target, wherein the traffic distribution strategy specifies a number of network coded packets to be routed over respective network paths of the set of network paths;
generate an NC group from a set of incoming packets, wherein the NC group comprises a set of network coded packets to achieve the reliability target; and
route individual subsets from the NC group to the respective network paths according to the traffic distribution strategy.
2. The edge compute node of claim 1 , wherein, to generate the NC group, the processor circuitry is to:
apply NC to packets in the set of incoming packets to produce the set of network coded packets in the NC group.
3. The edge compute node of claim 1 , wherein the processor circuitry is further to:
determine a code rate for encoding the set of incoming packets based on a ratio of the NC group size to a total number of network coded packets in the NC group.
4. The edge compute node of claim 1 , wherein the NC group size is a first NC group size among a plurality of NC group sizes, the traffic distribution strategy is a first traffic distribution strategy, and the processor circuitry is further to:
determine a respective traffic distribution strategy of a plurality of traffic distribution strategies for each NC group size of the plurality of NC group sizes based on the per path characteristics, a corresponding NC group size of the plurality of NC group sizes, and the reliability target.
5. The edge compute node of claim 4 , wherein the processor circuitry is further to:
perform header inspection on the set of packets to determine a packet identifier of the set of incoming packets associated with a reliability flow, wherein the packet identifier is one or more of a source node network address, a destination node network address, an ingress port number, an egress port number, and a quality of service (QOS) flow identifier.
6. The edge compute node of claim 5 , wherein the processor circuitry is further to:
perform automatic burst size detection based on a burst time spread parameter, wherein the burst time spread parameter specifies a maximum arrival time difference between a first packet of a burst of packets and a last packet of the burst of packets, wherein, to perform the automatic burst size detection, the processor circuitry is to:
determine a burst size by counting a number of incoming packets of the set of incoming packets arriving within a time window after arrival of the first packet of the burst of packets, wherein the time window has a same size as the burst time spread parameter;
select a traffic distribution strategy from the plurality of traffic distribution strategies based on the determined burst size; and
when the set of incoming packets are systematic code packets, route incoming packets of the set of incoming packets arriving within the time window for transmission over one or more network paths of the set of network paths without performing NC on the incoming packets arriving within the time window.
7. The edge compute node of claim 1 , wherein the per path characteristics include:
an erasure probability for transmitting packets over individual network paths of the set of network paths; and
delay characteristics indicating an expected time required for sending a number of packets over the individual network paths of the set of network paths.
8. The edge compute node of claim 7 , wherein the processor circuitry is further to:
operate an optimizer to determine an optimal subset of network paths from among the set of network paths and an optimal number of network coded packets to belong to the individual subsets of the NC group.
9. The edge compute node of claim 8 , wherein the processor circuitry is further to:
operate the optimizer to derive a worst-case delay for the individual network paths of the set of network paths based on the erasure probability, the delay characteristics, the reliability target, and the NC group size, wherein the derived worst-case delay is a maximum amount of acceptable delay for delivering the set of network coded packets over the individual network paths that still meets the reliability target;
determine a burst size based on the NC group size; and
generate a mapping function mapping the worst-case delay to the burst size, wherein the mapping function is for determining admission control criteria for performing admission control.
10. The edge compute node of claim 9 , wherein the processor circuitry is further to:
perform admission control for bursty-arrival traffic with a fixed burst size and a burst interval, wherein the burst interval is based on the worst-case delay to the burst size mapping;
perform admission control for periodic-arrival traffic based on an arrival period of the periodic-arrival traffic and the burst size; and
perform the admission control for the bursty-arrival traffic or the periodic-arrival traffic using a leaky-bucket traffic pacing mechanism or a token-bucket traffic pacing mechanism.
11. The edge compute node of claim 8 , wherein the processor circuitry is further to:
determine, based on the collected metrics, a delay bound for the individual network paths of the set of network paths; and
operate the optimizer to:
determine, as the optimal subset of network paths, a most resource efficient subset of the set of network paths among various other subsets of network paths from the set of network paths;
determine a feasibility for the set of network coded packets based on the per path characteristics, the NC group size, the reliability target, and the delay bound; and
generate a feasibility indicator based on the determined feasibility, wherein a value of the feasibility indicator indicates whether sending traffic over the individual network paths is feasible or not.
12. The edge compute node of claim 11 , wherein:
the feasibility indicator includes a value of ‘true’ to indicate that sending traffic over the individual network paths is feasible, and the feasibility indicator includes a value of ‘false’ to indicate that sending traffic over the individual network paths is not feasible; or
the feasibility indicator includes a numeric value, wherein the numeric value being above a predetermined or configured threshold indicates that sending traffic over the individual network paths is feasible, and the numeric value being below the predetermined or configured threshold indicates that sending traffic over the individual network paths is not feasible.
13. The edge compute node of claim 11 , wherein the processor circuitry is further to:
perform admission control using the feasibility indicator, wherein performance of the admission control includes:
rejection of admission of the set of network coded packets when the value of the feasibility indicator indicates that sending traffic over the individual network paths is not feasible; and
admission of the set of network coded packets when the value of the feasibility indicator indicates that sending traffic over the individual network paths is feasible.
14. The edge compute node of claim 13 , wherein the processor circuitry is further to:
determine whether delivering individual subsets of the optimal subsets of network coded packets to be sent over respective ones of the optimal subset of network paths exceeds a reliability flow resource allocation threshold during a period that is within an inter-burst interval or the delay bound; and
reject admission of the set of network coded packets when delivering at least one of the optimal subsets exceeds the reliability flow resource allocation threshold during the period.
15. The edge compute node of claim 1 , wherein the one or more data sources include the set of NANs or the one or more user devices, and the metrics related to the multi-access communication network include access network measurements, wherein the access network measurements include one or more of reference signal received power (RSRP), reference signal received quality (RSRQ), received signal strength indicator (RSSI), received channel power indicator (RCPI), received signal to noise indicator (RSNI), average noise plus interference (ANPI), signal-to-noise ratio (SNR), signal-to-noise and interference ratio (SINR), expected peak data rate, end-to-end delay measurements, backhaul delay measurements, scheduling delay measurements, packet drop rate, and packet error ratio.
16. The edge compute node of claim 15 , wherein the metrics related to the multi-access environment further include capabilities indicated by one or more configurations, wherein the capabilities include one or more of QoS capabilities included in a QoS configuration, wireless local area network (WLAN) capabilities included in a WLAN configuration, and radio resource control (RRC) information included in a RRC configuration.
17. The edge compute node of claim 16 , wherein the metrics related to the multi-access environment further include application layer metrics, wherein the application layer metrics include one or more of priority indicators for data flows, QoS requirements of the data flows, wherein the QoS requirements include one or more of the reliability target, a latency bound, a minimum guaranteed bit rate, and a maximum burst size.
18. The edge compute node of claim 17 , wherein the metrics related to the multi-access environment further include data plane telemetry metrics, wherein the data plane telemetry metrics include one or more of traffic volume estimates of reliability flows, traffic volume estimates of non-reliability flows, flow path and latency metrics, packet drop metrics, traffic congestion metrics, platform load metrics, and application load metrics.
19. One or more non-transitory computer readable media (NTCRM) comprising instructions for operating a reliability enhancement engine (REE), wherein execution of the instructions by one or more processors of a compute node is to cause the compute node to operate the REE to:
collect metrics related to a multi-access communication network from one or more data sources within the multi-access communication network, wherein the one or more data sources include a set of network access nodes (NANs) in the multi-access communication network or one or more user devices connected to the set of NANs;
determine, based on the collected metrics, a reliability target, a network coding (NC) group size, and per path characteristics for each network path of a set of network paths in the multi-access communication network;
determine a traffic distribution strategy based on the per path characteristics, the NC group size, and the reliability target;
generate an NC group from a set of incoming packets, wherein the NC group comprises a set of network coded packets to achieve the reliability target; and
route individual subsets from the NC group to respective network paths of the set of network paths according to the traffic distribution strategy.
20. The one or more NTCRM of claim 19 , wherein execution of the instructions is to cause the compute node to operate the REE to:
determine a number of available transmission units on each network path of the set of network paths;
perform NC on the set of incoming packets to obtain the set of network coded packets; and
deliver the individual subsets of the set of network coded packets to the respective network paths of the set of network paths.
21. The one or more NTCRM of claim 19 , wherein, to determine the number of available transmission units on each network path, execution of the instructions is to cause the compute node to operate the REE to:
calculate a number of independently transmitted packets supported on each network path within a transmission delay bound of the set of incoming packets.
22. The one or more NTCRM of claim 19 , wherein, to perform NC on the set of incoming packets and deliver the individual subsets to the respective network paths, execution of the instructions is to cause the compute node to operate the REE to:
group a succession of the set of incoming packets, wherein a length of the succession of the set of incoming packets is specified by a predefined NC configuration;
determine a number of network coded packets in the set of network coded packets to be generated from each grouping of the incoming packets;
determine respective sizes of corresponding subsets of the individual subsets to be delivered to the respective network paths;
generate the individual subsets from each grouping of the incoming packets according to the NC configuration and a total number of network coded packets in the set of network coded packets; and
deliver the individual subsets to the respective network paths according to the respective sizes.
23. The one or more NTCRM of claim 22 , wherein, to determine the respective sizes of the corresponding subsets, execution of the instructions is to cause the compute node to operate the REE to:
operate an optimizer to determine the respective sizes of the corresponding subsets based on the NC group size, the reliability target, an erasure probability of the respective network paths, a spectral efficiency of the respective network paths, and a number of available transmission units.
24. The one or more NTCRM of claim 23 , wherein execution of the instructions is to cause the compute node to operate the REE to:
determine a score for each network path of the set of network paths based on the erasure probability of a corresponding one of each network path, a spectral efficiency of the corresponding one of each network path, the NC group size, and the reliability target;
generate a preference order for each of the respective network paths according to the score of each network path; and
operate the optimizer to determine the respective sizes of the corresponding subsets further based on the preference order.
25. The one or more NTCRM of claim 24 , wherein execution of the instructions is to cause the compute node to operate the REE to:
operate the optimizer to determine the respective sizes of the corresponding subsets according to a greedy-type algorithm using the preference order, wherein the greedy-type algorithm is a line-search greedy algorithm or a binary-search greedy algorithm; and
apply a recursive final-unit adjustment procedure to a result of the greedy-type algorithm.
26. The one or more NTCRM of claim 19 , wherein the REE is part of a multi-access traffic manager (MA-TM), and the compute node is an edge compute node.
27. The one or more NTCRM of claim 26 , wherein:
the MA-TM is part of an edge platform of an edge compute node, and the REE is an edge app operated by the edge compute node; or
the REE and the MA-TM are implemented as respective edge applications (apps) of an edge compute node.
28. The one or more NTCRM of claim 26 , wherein:
the edge compute node is a Multi-Access Edge Computing (MEC) server of a MEC framework, and the REE is implemented as a MEC app configured to be operated by a MEC platform of the MEC server;
the edge compute node is a Multi-Access Management Services (MAMS) server or a MAMS client of a MAMS framework, and the REE is implemented as a Network Connection Manager (NCM) configured to be operated by the MAMS server or a Client Connection Manager (CCM) configured to be operated by the MAMS client;
the edge compute node is a Radio Access Network (RAN) intelligent controller (RIC) of an Open RAN Alliance (O-RAN) framework, and the REE is implemented as an xApp configured to be operated by the RIC;
the edge compute node is a central unit (CU) of a CU/distributed unit (DU)-split architecture of a next generation (NG)-RAN architecture of a third generation partnership project (3GPP) cellular network, and the REE is implemented as a network function (NF) configured to be operated by the CU; or
the edge compute node is an Edge Enabler Server (EES) or an Edge Configuration Server (ECS) of a 3GPP edge computing architecture, and the REE is implemented as an Edge Application Servers (EAS) configured to be operated by the EES or the ECS.