To configure the UPF ingress interface type, use the following configuration:

To enable the upf-ingress command, you must configure the require upf command on the UPF. You need the UPF license to enable the require upf command.

configure 
   context context_name 
      user-plane-service service_name 
         [ no ] associate gtpu-service gtpu_service_name upf-ingress interface-type [ n9 | s5u | s8u | n3  | n9-s5u-s8u  ] 
         end 

NOTES:

• associate gtpu-service gtpu_service_name : Associate the GTP-U service with the user plane service.

• upf-ingress : Configure the interface type as UPF ingress.

• interface-type [ n9 | s5u | s8u | n3 | n9-s5u-s8u ] : Configure the desired GTP-U ingress interface type.

• If the N3 GTP-U ingress service is not present, UPF uses the default interface.

• If the configuration for N9, S5u, or S8u interfaces is not present, UPF performs the following actions:

  • UPF looks to use the N9 interface first followed by the S5u and S8u interfaces.

  • If there is no GTP-U service, UPF uses the default interface.

The echo response in a GTP-U service provides the availability of peer UPF.

The following is a sample echo response configuration for the N9 interface in hUPF:

gtpu-service n9-ingress
   echo-interval 60
   echo-retransmission-timeout 1
   bind ipv4-address 60.60.41.5 bearer-type all
   exit

The following is a sample echo response configuration for the N9 interface in vUPF:

gtpu-service n9-egress
   echo-interval 60
   echo-retransmission-timeout 1
   bind ipv4-address 60.60.42.5 bearer-type all
   exit

The following is a sample configuration to separate the N3 and N9 BGP routes for hUPF.

 config
  context SAEGW
    ip vrf mpls-vrf-1
    #exit
    ip vrf mpls-vrf-2
    #exit
    mpls bgp forwarding
    bfd-protocol
    #exit
    ip prefix-list name up seq 15 permit 50.50.17.4/32
    ip prefix-list name up seq 20 permit 50.50.17.5/32
    ip prefix-list name up seq 25 permit 50.50.17.6/32
    ip prefix-list name up seq 40 permit 50.50.17.7/32
    ip prefix-list name n9-up1 seq 50 permit 60.60.17.5/32
    ip prefix-list name n9-up1 seq 55 permit 60.60.17.6/32
    route-map up_routes1 permit 10
      match ip address prefix-list n9-up1
    #exit
    route-map up_routes permit 10
      match ip address prefix-list up
    #exit
    router bgp 61435
      maximum-paths ebgp 4
      neighbor 20.14.35.100 remote-as 65000
      neighbor 20.14.35.100 ebgp-multihop
      neighbor 20.14.35.100 update-source 20.14.35.98
      neighbor 2001:4888:121:1435::1 remote-as 65000
      neighbor 2001:4888:121:1435::1 ebgp-multihop
      neighbor 2001:4888:121:1435::1 update-source 2001:4888:121:1435::2
      neighbor 20.14.40.100 remote-as 65000
      neighbor 20.14.40.100 ebgp-multihop
      neighbor 20.14.40.100 update-source 20.14.40.98
      address-family ipv4
        neighbor 20.14.35.100 max-prefix 8192 threshold 50
        neighbor 20.14.35.100 route-map up_routes out
        neighbor 20.14.35.100 send-community extended
        neighbor 20.14.40.100 max-prefix 8192
        neighbor 20.14.40.100 route-map up_routes1 out
        neighbor 20.14.40.100 send-community extended
        redistribute connected
        redistribute static
      #exit
      address-family ipv6
        neighbor 2001:4888:121:1435::1 activate
        redistribute connected
        redistribute static
      #exit
    #exit
    interface SxC loopback
      ip address 50.50.17.4 255.255.255.255
    #exit
    interface SxC-ipv6 loopback
      ipv6 address 2001:4888:50:50::17/128
    #exit
    interface gn-1
      ip address 20.14.35.98 255.255.255.0
      ipv6 address 2001:4888:121:1435::2/64 secondary
    #exit
    interface gn-2
      ip address 20.14.36.98 255.255.255.0
      ipv6 address 2001:4888:121:1436::2/64 secondary
    #exit
    interface gn-n9-ingress
      ip address 20.14.40.98 255.255.255.0
    #exit
    interface n9-ingress-hupf loopback
      ip address 60.60.17.5 255.255.255.255
    #exit
    interface pgw-gtpu loopback
      ip address 50.50.17.5 255.255.255.255
    #exit
    interface sgw-gtpu-egress loopback
      ip address 50.50.17.6 255.255.255.255
    #exit
    interface sgw-gtpu-ingress-phazr loopback
      ip address 60.60.17.6 255.255.255.255
    #exit
    subscriber default
    exit
    aaa group default
    #exit
    gtpp group default
      gtpp limit-secondary-rat-usage 32
    #exit
    gtpu-service SxC
      echo-interval 60
      echo-retransmission-timeout 1
      bind ipv4-address 50.50.17.4 ipv6-address 2001:4888:50:50::17 bearer-type all
    exit
    gtpu-service n9-ingress
      bind ipv4-address 60.60.17.5 bearer-type all
    exit
    gtpu-service pgw-gtpu
      bind ipv4-address 50.50.17.5 bearer-type all
    exit
    gtpu-service sgw-gtpu-egress
      bind ipv4-address 50.50.17.6 bearer-type all
    exit
    gtpu-service sgw-gtpu-ingress-phazr
      echo-interval 60
      echo-retransmission-timeout 1
      bind ipv4-address 60.60.17.6 bearer-type all
    exit
    sx-service sx-svc
      instance-type userplane
      bind ipv4-address 50.50.17.4 ipv6-address 2001:4888:50:50::17
      sxa max-retransmissions 5
      sxa retransmission-timeout-ms 1000
      sxb max-retransmissions 5
      sxb retransmission-timeout-ms 1000
      sxab max-retransmissions 5
      sxab retransmission-timeout-ms 1000
      sx-protocol heartbeat retransmission-timeout 15
      sx-protocol heartbeat max-retransmissions 3
      sx-protocol association debug-mode debug-reattempt-timeout 1
      no sx-protocol compression
    exit
    user-plane-service user_plane_svc
      associate gtpu-service pgw-gtpu upf-ingress
      associate gtpu-service n9-ingress upf-ingress interface-type n9
      associate gtpu-service sgw-gtpu-ingress-phazr sgw-ingress
      associate gtpu-service sgw-gtpu-egress sgw-egress
      associate gtpu-service SxC cp-tunnel
      associate sx-service sx-svc
      associate fast-path service
      associate control-plane-group SAEGW
      associate userplane-load-control-profile LCP
      associate userplane-overload-control-profile OLCP
    exit

The following is a sample configuration to separate the N3 and N9 BGP routes for vUPF.

context SAEGW
    router bgp 61441
      neighbor 20.14.41.100 remote-as 65000
      neighbor 20.14.41.100 ebgp-multihop
      neighbor 20.14.41.100 update-source 20.14.41.98
      neighbor 2001:4888:121:1441::1 remote-as 65000
      neighbor 2001:4888:121:1441::1 ebgp-multihop
      neighbor 2001:4888:121:1441::1 update-source 2001:4888:121:1441::2
      address-family ipv4
        redistribute connected
        redistribute static
      #exit
      address-family ipv6
        neighbor 2001:4888:121:1441::1 activate
        redistribute connected
        redistribute static
      #exit
    #exit
    interface SxC loopback
      ip address 50.50.18.4 255.255.255.255
    #exit
    interface SxC-ipv6 loopback
      ipv6 address 2001:4888:50:50::18/128
    #exit
    interface gn-1
      ip address 20.14.41.98 255.255.255.0
      ipv6 address 2001:4888:121:1441::2/64 secondary
    #exit
    interface gn-2
      ip address 20.14.42.98 255.255.255.0
      ipv6 address 2001:4888:121:1442::2/64 secondary
    #exit
    interface n9-egress loopback
      ip address 50.50.18.8 255.255.255.255
    #exit
    interface pgw-gtpu loopback
      ip address 50.50.18.5 255.255.255.255
    #exit
    interface pgw-gtpu-ipv6 loopback
      ipv6 address 2001:4888:50:50::18:5/128
    #exit
    interface sgw-gtpu-egress loopback
      ip address 50.50.18.6 255.255.255.255
    #exit
    interface sgw-gtpu-egress-ipv6 loopback
      ipv6 address 2001:4888:50:50::18:6/128
    #exit
    interface sgw-gtpu-ingress-phazr loopback
      ip address 50.50.18.7 255.255.255.255
    #exit
    interface sgw-gtpu-ingress-phazr-ipv6 loopback
      ipv6 address 2001:4888:50:50::18:7/128
    #exit
    subscriber default
    exit
    aaa group default
    #exit
    gtpp group default
      gtpp limit-secondary-rat-usage 32
    #exit
    gtpu-service SxC
      echo-interval 60
      echo-retransmission-timeout 1
      bind ipv4-address 50.50.18.4 ipv6-address 2001:4888:50:50::18 bearer-type all
    exit
    gtpu-service n9-egress
      bind ipv4-address 50.50.18.8 bearer-type all
    exit
    gtpu-service pgw-gtpu
      bind ipv4-address 50.50.18.5 ipv6-address 2001:4888:50:50::18:5 bearer-type all
    exit
    gtpu-service sgw-gtpu-egress
      bind ipv4-address 50.50.18.6 ipv6-address 2001:4888:50:50::18:6 bearer-type all
    exit
    gtpu-service sgw-gtpu-ingress-phazr
      echo-interval 60
      echo-retransmission-timeout 1
      bind ipv4-address 50.50.18.7 ipv6-address 2001:4888:50:50::18:7 bearer-type all
    exit
    sx-service sx-svc
      instance-type userplane
      bind ipv4-address 50.50.18.4 ipv6-address 2001:4888:50:50::18
      sxa max-retransmissions 5
      sxa retransmission-timeout-ms 1000
      sxb max-retransmissions 5
      sxb retransmission-timeout-ms 1000
      sxab max-retransmissions 5
      sxab retransmission-timeout-ms 1000
      sx-protocol heartbeat retransmission-timeout 15
      sx-protocol heartbeat max-retransmissions 3
      sx-protocol association debug-mode debug-reattempt-timeout 1
      no sx-protocol compression
    exit
    user-plane-service user_plane_svc
      associate gtpu-service pgw-gtpu upf-ingress
      associate gtpu-service n9-egress upf-egress
      associate gtpu-service sgw-gtpu-ingress-phazr sgw-ingress
      associate gtpu-service sgw-gtpu-egress sgw-egress
      associate gtpu-service SxC cp-tunnel
      associate sx-service sx-svc
      associate fast-path service
      associate control-plane-group SAEGW
      associate userplane-load-control-profile LCP
      associate userplane-overload-control-profile OLCP
    exit
    ip route 0.0.0.0 0.0.0.0 20.14.41.100 gn-1
    ip igmp profile default
    #exit
  #exit 

Note	For RCM UPF, add all the IP prefix-lists in day 0.5 configuration.

To verify the configuration, use the following command:

• show user-plane-service all

In the Private 5G networks, where the uplink traffic is more than the downlink traffic, the number of gNodeBs are limited. UPF achieves the optimal distribution of subscriber traffic using the Receive Side Scaling (RSS) concept. However, due to limited number of VPP worker threads on UPF, the ingress traffic does not get evenly distributed across them. It limits the UPF from optimally utilizing all the worker threads, and therefore impacts the subscriber’s uplink throughput.

UPF allows utilizing the VPP worker threads in an optimal manner by creating the unique GTPU 3-tuple hash entries on the UPF’s forwarding plane. It ensures an optimum distribution of ingress traffic on the N3 interface. It enables UPF to perform traffic processing to its maximum capacity ensuring high uplink throughput for Private 5G networks.

The even distribution of ingress traffic across all the available worker threads on the UPF is ensured by creating the unique 3-tuple hash entries.

The GTPU header source IP is the gNodeB’s N3 service egress IP, while the source port is the standard port in the absence of port randomization support on the gNodeB. With these two values being constants per gNodeB, the only variable that can be achieved in the GTPU header is the destination IP (N3 service ingress IP address).

The following points describe how the unique 3-tuple hash entries are created:

• GTPU addresses are assigned in FTEID in Sx Session Establish Response for a subscriber session in a round-robin manner. However, this round-robin mechanism is ensured only within the session manager.

• When a session manager is selected, the GTPU address in FTEID is allocated in a round-robin manner to all the sessions landing on the selected session manager.

• Deletion of IP address from the GTPU service or the interface does not cause the GTPU service to stop. The existing calls will be deleted. However, the uplink traffic of the subscriber to which the deleted IP address was assigned will not be received at UPF.

• If every IP address is configured under a separate interface, the interface must be configured as a loopback (multiple non-loopback interfaces cannot share the same subnet). If any interface goes down and comes alive again, the GTPU service does not stop as the interfaces are configured as loopback.

• Every address pair is one GTPU endpoint for a GTPU service. Every endpoint should belong to the same subnet.

Note	For optimum performance, the number of GTPU bind addresses must be equal to the number of VPP worker threads. A maximum of 12 IP address pairs (either IPv4 or IPv6 or IPv4+IPv6) can be configured per GTPU service.

To configure multiple bind addresses on the N3 interface, use the following sample configuration:

config 
   context context_name 
      gtpu-service service_name 
         bind { ipv4-address ipv4_address | ipv6-address ipv6_address [ bearer-type { non-ims-media | ims-media | all } ] } 
         exit 

NOTES:

• bind { ipv4-address ipv4_address | ipv6-address ipv6_address } —Bind the GTPU service to the IPv4 address or IPv6 address on the N3 interface.

• [ bearer-type { non-ims-media | ims-media | all } ] —Specify the type of bearer to be associated with the bind address. bearer-type allows three types of media traffic:

  • non-ims-media —Configure bind address for non-ims media only.

  • ims-media —Configure bind address for ims-media traffic only.

  • all —Configure bind address to handle all types of bearer traffic. This is the default behavior.

Note	If GTPU echo is enabled in GTPU service, the Echo request will be started with the first bind address configured in GTPU service. The bind address configuration is stored in the sorted list. As a result, the GTPU echo starts with the numerically lowest bind IP address.

This section describes operations, administration, and maintenance support for this feature.