IOS-XR

This section contains notes on IOS-XR router configs and network design.

Segment Routing

1. SRGB: we use a custom label block in our lab. The default is 16000 - 23999.

   segment-routing
    global-block 100000 163999   
   

2. Enable SR in ISIS

   router isis 100
    is-type level-2-only
    net 49.0901.0000.0000.0006.00
    address-family ipv4 unicast
     metric-style wide
     advertise link attributes
     mpls traffic-eng level-2-only
     mpls traffic-eng router-id Loopback0
     maximum-paths 32
     segment-routing mpls
    !
    interface Loopback0
     passive
     address-family ipv4 unicast
      prefix-sid absolute 100006 #(1)!
     !
   

   1. This can also be an SRGB index value

3. Enable distribution of SR Prefix-SID information in BGP-LU (Optional, but very valuable for multi-domain networks)

   route-policy SID($SID)
     set label-index $SID
   end-policy
   !
   router bgp 100000
   !
    address-family ipv4 unicast
     network 10.0.0.1/32 route-policy SID(1) #(1)! 
     allocate-label all
    !
    neighbor 10.1.1.0   #(2)!
     remote-as 65000
     !       
     address-family ipv4 labeled-unicast
      route-policy pass in
      route-policy pass out
   !
   

   1. BGP Prefix-SID index 1, results in SR label value 100001 based on our SRGB
   2. ASBR

Setting up BMP and BGP-LS

Diagram

BGP-LS

1. Setup route-reflectors to receive LS messages from clients, but to not pass LS messages back out:

   BGP-LS specific information shown:
   
   router bgp 100000
    !
    address-family link-state link-state
    !
    neighbor <neighbor IP>
     address-family link-state link-state
      route-policy pass in
      route-policy drop out
   

Note

The RR's technically only need a copy of BGP-LS/LSDB from one router.

1. Configure one or more RR clients to pass LS messages to the RR's.

   BGP-LS specific information shown:

   router isis 100
    distribute link-state level 2 #(1)!
    !
    address-family ipv4 unicast
     advertise link attributes #(2)!
    !
   router bgp 100000
    !
    address-family link-state link-state
    !
    neighbor 192.0.2.50 #(3)!
     !
     address-family link-state link-state
      route-policy drop in
      route-policy pass out
   

   1. Distribute my LSDB into local BGP-LS
   2. All ISIS nodes should have this line as it adds their SR TLVs into the domain's LSDB
   3. Route Reflector

2. Add Egress Peer Engineering data to BGP-LS feed (Optional)

   On an ASBR node, or Internet peering node add the following (it is assumed v4/v6 AFIs are already enabled):

   router bgp 100000
    !
    address-family link-state link-state
    !
    neighbor 192.0.2.50 #(1)!
     !       
     address-family link-state link-state
      route-policy pass out
      !
     !
    neighbor 203.0.113.50 #(2)!
     remote-as 64496 
     egress-engineering
    !
   

   1. Route Reflector
   2. External Peer

BMP (BGP Monitoring Protocol)

We collect topology data with BMP as it provides data from multiple AFI/SAFI combinations (not just BGP-LS) While we anticipate most operators to run a BGP-free core, we'll generally want to collect BMP messages from route-reflectors and all BGP speakers with external facing peering sessions (ASBRs, Peering, etc.):

1. Configure BMP Server on RR's and on ASBRs and peering routers:

   bmp server 1
    host 10.0.250.2 port 30511
    description Jalapeno GoBMP 
    update-source Loopback0 #(1)!
    flapping-delay 60
    initial-delay 5
    stats-reporting-period 60
    initial-refresh delay 30 spread 2
   

   1. Alternatively, update-source MgmtEth0/RP0/CPU0/0

2. Configure export of BMP data for BGP messages/advertisements from specific peers:

   1. Route reflector

      router bgp 100000
      neighbor <neighbor IP> #(1)!
      bmp-activate server 1
      

      1. Assuming clients are VPNv4/v6 PE's, this also captures VPNv4/6 messages

   2. ASBR/Peering

       router bgp 100000
       !
       neighbor 192.0.2.100
        remote-as 64500
        description External Peer 72
        egress-engineering
        bmp-activate server 1
      

Streaming Telemetry (Model Driven Telemetry)

The following configuration snippet provides a reference for sending streaming telemetry to Jalapeno.

telemetry model-driven
destination-group jalapeno
    vrf <name> // optional 
    address-family ipv4 192.0.2.10 port 32400
    encoding self-describing-gpb
    protocol grpc no-tls
    !
!
sensor-group cisco_models 
    sensor-path Cisco-IOS-XR-pfi-im-cmd-oper:interfaces/interface-xr/interface #(1)!
    sensor-path Cisco-IOS-XR-infra-tc-oper:traffic-collector/afs/af/counters/prefixes/prefix #(2)!
    sensor-path Cisco-IOS-XR-fib-common-oper:mpls-forwarding/nodes/node/label-fib/forwarding-details/forwarding-detail #(3)!
!
sensor-group openconfig_interfaces
    sensor-path openconfig-interfaces:interfaces/interface #(4)!
!
subscription base_metrics
    sensor-group-id cisco_models sample-interval 10000
    sensor-group-id openconfig_interfaces sample-interval 10000
    destination-id jalapeno
    source-interface MgmtEth0/RP0/CPU0/0
!

1. Interface statistics
2. SR traffic collector statistics
3. Per-MPLS label forwarding statistics
4. Openconfig interface statistics