In the last videos we mainly focused on explicit path capability of the segment routing traffic engineering, but path can be dynamically selected by routers themselves based on the lowest metric and taking into account any constraint, if any exist. Segment Routing Traffic Engineering Dynamic path is the topic of this video.

Some of the constraints that you can configure are including or excluding a specific IP in the dynamic path, including or excluding links with a specific tag or affinity, including or excluding links with specific shared risk link group, paths with a metric less than the maximum pre-defined value , or paths that have no more SID than the maximum predefined value.

Segment Routing Traffic Enginnering Explicit Path Configuration

Segment Routing Traffic Engineering Dynamic Path Topology

segment routing traffic engineering dynamic path topology
segment routing traffic engineering dynamic path topology

As an example in this section we will define some constraints based on link tag or affinity. With this method we can mark a link as GOLD, SILVER, RED, BLUE and whatever which has a meaning to our infrastructure. Then we can ask the router to find a path that includes or excludes certain tags. With segment routing traffic engineering, routers have the ability to dynamically choose the path based on the lowest metrics and considering any constraint.

In this topology we have three PE routers, PE1, PE2 and PE3 with the IP addresses 1.1.1.1, 2.2.2.2 and 3.3.3.3 and two P routers, P1 and P2 with the IP addresses 11.11.11.11 and 12.12 .12.12. PE routers, of course, can communicate with each other through two routers, P1 or P2 router. There is also one link between P1 and P2 router in our topology.

metric of all links are 10 by default in our topology. We change the metric of the two links PE1-P1 and P2-PE2 to the value 25. So the best path from PE1 to PE2 is through PE1-P2-P1-PE2, if we do not consider any other constraint to our PE1 router.

For a better understanding, let’s assume that we mark the links of the network connected to P1 with the GOLD affinity and the links of the network connected to P2 with the SILVER affinity. We also mark the connection between P1-PE2 with the RED affinity and We do not mark the connection between P1-P2. As you noticed, we can mark a link with more than one affinity. The link between P1-PE2 has two affinities, GOLD and RED.

Now we want to write two Policies from PE1 to PE2. One for SILVER customers, in which we ask router to dynamically choose the best path with the condition that exclude any link with GOLD affinity. We write also other policy for GOLD customers in which the constraint is to exclude any link with RED affinity.

As you know, the best path from PE1 to PE2 without considering any constraint is PE1-P2-P1-PE2, but the GOLD and SILVER policies choose a different path because this path does not meet the constraint. In GOLD politics we are not allowed to choose a path with RED affinity. In the SILVER policy, we are not allowed to choose a path with a GOLD affinity. Since the connection between P1-PE2 has both tags, it cannot choose this path, instead the traffic of both customers goes through PE1-P2-PE2.

Segment Routing Traffic Engineering Reference

Segment Routing Traffic Engineering Dynamic Path Configuration

For the implementation of Segment Routing Traffic Engineering Dynamic Path from PE1 to PE2, I have shown here the configuration of PE1 and P1 here as an example.

segment routing traffic engineering dynamic path
segment routing traffic engineering dynamic path

The configuration of PE1 is divided in three parts.  The ISIS IGP protocol with the address family IPv4 is activated in all interfaces. The point-to-point network type is enabled in Ethernet interfaces to ignore the DIS selection. The metric of all interfaces is 10 by default. But the metric of GE0/0/0/0 is changed to 25.

!!! PE1
mpls traffic-eng
!
mpls oam
!
router isis LAB
 is-type level-2-only
 net 49.0001.0000.0000.0001.00
 distribute link-state
 address-family ipv4 unicast
  metric-style wide
  mpls traffic-eng level-2-only
  mpls traffic-eng router-id Loopback0
  segment-routing mpls sr-prefer
 !
 interface Loopback0
  address-family ipv4 unicast
   prefix-sid absolute 16001
  !
 !
 interface GigabitEthernet0/0/0/0
  point-to-point
  address-family ipv4 unicast
   metric 25
  !
 !
 interface GigabitEthernet0/0/0/1
  point-to-point
  address-family ipv4 unicast
!
segment-routing
 traffic-eng
  interface GigabitEthernet0/0/0/0
   affinity
    name GOLD
   !
  !
  interface GigabitEthernet0/0/0/1
   affinity
    name SILVER
   !
  !
  policy GOLD
   color 10 end-point ipv4 2.2.2.2
   candidate-paths
    preference 100
     dynamic
      metric
       type igp
      !
     !
     constraints
      affinity
       exclude-any
        name RED
       !
      !
     !
    !
   !
  !
  policy SILVER
   color 20 end-point ipv4 2.2.2.2
   candidate-paths
    preference 100
     dynamic
      metric
       type igp
      !
     !
     constraints
      affinity
       exclude-any
        name GOLD
       !
      !
     !    
    !
   !
  !
  affinity-map
   name RED bit-position 2
   name GOLD bit-position 0
   name SILVER bit-position 1

“mpls traffic-eng” is activated globally and also activated in ISIS context mode with two commands “mpls traffic-eng level-2-only”, since ISIS itself is activated in level-2 only and “mpls traffic-eng router-id loopback0”.

segment routing is also activated in the ISIS context mode with the command “segment-routing mpls”. Prefix SID of 16001 is configured in In Loopback 0 interface.

The “distribute link-state” command is required in the PE1 router that wants to find a dynamic path for segment routing. with this command, all links information including the affinity is distributed in the PE1 router database

The “mpls oam” command is only used for ping and traceroute in the MPLS network for better troubleshooting which is configured in all routers but is not required.

In the second part, I configured segment routing traffic engineering with dynamic policy for both GOLD and SILVER customers and for the destination PE2 with IP address 2.2.2.2. for PE3, similar policies can be configured which I have not configured.

Both policies are to route traffic on the best path with the lowest IGP metric. in addition to IGP metric, We also have the option of TE metric and also Delay for choosing the best path. by default, TE metric is the same as IGP metric which can be changed by configuration.

But the constraint is different for GOLD and SILVER customers. for GOLD customers, we need to ensure that traffic does not go through links with RED affinity. For SILVER customers, we need to ensure that traffic does not go through links with GOLD affinity.

In the third part of the PE1 configuration, we mark the links based on our definition. In PE1 GE0/0/0/0 is marked with GOLD affinity and GE0/0/0/1 with SILVER affinity. How is the marking mapped in the corresponding field in the ISIS database? The first bit is to be set if a link is GOLD. The second bit is for SILVER links and the third bit is for links with RED marking.

The configuration of P1 router is also displayed here. The configuration of ISIS in P1 is the same as PE1. I do not explain more. But we do not need to configure “distribute link-state” command in P1 since we do not want to configure traffic engineering policy in P1 router. The prefix SID of 16011 is configured for the loopback 0 interface of the P1 router since the IP address of the loopback0 interface is 11.11.11.11.
I also marked the links in P1 routers. GE0/0/0/0 is not marked at all. GE0/0/0/1 and GE0/0/0/3 are marked with GOLD affinity. GE0/0/0/2, which is connected to PE2, is marked with both RED and GOLD affinity.

!!! P1
mpls traffic-eng
!
mpls oam
!
router isis LAB
 is-type level-2-only
 net 49.0001.0000.0000.0011.00
 address-family ipv4 unicast
  metric-style wide
  mpls traffic-eng level-2-only
  mpls traffic-eng router-id Loopback0
  segment-routing mpls sr-prefer
 !
 interface Loopback0
  address-family ipv4 unicast
   prefix-sid absolute 16011
  !
 !
 interface GigabitEthernet0/0/0/0
  point-to-point
  address-family ipv4 unicast
  !
 !
 interface GigabitEthernet0/0/0/1
  point-to-point
  address-family ipv4 unicast
   metric 25
  !       
 !
 interface GigabitEthernet0/0/0/2
  point-to-point
  address-family ipv4 unicast
  !
 !
 interface GigabitEthernet0/0/0/3
  point-to-point
  address-family ipv4 unicast
  !
 !
!
segment-routing
 traffic-eng
  interface GigabitEthernet0/0/0/0
  !
  interface GigabitEthernet0/0/0/1
   affinity
    name GOLD
   !
  !
  interface GigabitEthernet0/0/0/2
   affinity
    name RED
    name GOLD
   !
  !
  interface GigabitEthernet0/0/0/3
   affinity
    name GOLD
   !
  !
  affinity-map
   name RED bit-position 2
   name GOLD bit-position 0
   name SILVER bit-position 1
  !
 !
!

Let’s now check the result of the configuration

First we traceroute from PE1 to PE2 to check the best IGP path. as you can see it goes through PE1-P1-P2-PE2.

RP/0/RP0/CPU0:PE1# traceroute 2.2.2.2
Tue Aug 24 10:54:54.048 UTC

Type escape sequence to abort.
Tracing the route to 2.2.2.2

 1  10.1.12.12 [MPLS: Label 16002 Exp 0] 111 msec  96 msec  106 msec 
 2  10.11.12.11 [MPLS: Label 16002 Exp 0] 90 msec  86 msec  84 msec 
 3  10.2.11.2 86 msec 
 *  97 msec 

Now let’s check the path chosen dynamically by the policies configured in PE1.

RP/0/RP0/CPU0:PE1#show segment-routing traffic-eng policy 
Tue Aug 24 10:55:19.516 UTC

SR-TE policy database
---------------------

Color: 10, End-point: 2.2.2.2
  Name: srte_c_10_ep_2.2.2.2
  Status:
    Admin: up  Operational: up for 01:06:55 (since Aug 24 09:48:24.408)
  Candidate-paths:
    Preference: 100 (configuration) (active)
      Name: GOLD
      Requested BSID: dynamic
      Constraints:
        Affinity:
          exclude-any:
          RED
      Dynamic (valid)
        Metric Type: IGP,   Path Accumulated Metric: 35 
          16012 [Prefix-SID, 12.12.12.12]
          24003 [Adjacency-SID, 10.2.12.12 - 10.2.12.2]
  Attributes:
    Binding SID: 24004
    Forward Class: Not Configured
    Steering BGP disabled: no
    IPv6 caps enable: yes

Color: 20, End-point: 2.2.2.2
  Name: srte_c_20_ep_2.2.2.2
  Status:
    Admin: up  Operational: up for 01:06:55 (since Aug 24 09:48:24.408)
  Candidate-paths:
    Preference: 100 (configuration) (active)
      Name: SILVER
      Requested BSID: dynamic
      Constraints:
        Affinity:
          exclude-any:
          GOLD
      Dynamic (valid)
        Metric Type: IGP,   Path Accumulated Metric: 35 
          16012 [Prefix-SID, 12.12.12.12]
          24003 [Adjacency-SID, 10.2.12.12 - 10.2.12.2]
  Attributes:
    Binding SID: 24005
    Forward Class: Not Configured
    Steering BGP disabled: no
    IPv6 caps enable: yes

As you can see we have two policies. The result of dynamic path in both policies are different from IGP best path. first prefix-SID 16012 which points to P2 and then adjacency-SID 24003 which points to the link between P2 and PE2 and you can check with command “show isis adjacency detail”

RP/0/RP0/CPU0:P2#show isis adjacency detail 
Wed Aug 25 16:07:58.877 UTC

IS-IS LAB Level-2 adjacencies:
System Id      Interface                SNPA           State Hold Changed  NSF IPv4 IPv6
                                                                               BFD  BFD 
PE1            Gi0/0/0/1                *PtoP*         Up    29   1d06h    Yes None None
  Area Address:           49.0001
  Neighbor IPv4 Address:  10.1.12.1*
  Adjacency SID:          24004
  Non-FRR Adjacency SID:  24005
  Topology:               IPv4 Unicast
PE2            Gi0/0/0/2                *PtoP*         Up    28   2d00h    Yes None None
  Area Address:           49.0001
  Neighbor IPv4 Address:  10.2.12.2*
  Adjacency SID:          24002
  Non-FRR Adjacency SID:  24003
  Topology:               IPv4 Unicast
PE3            Gi0/0/0/3                *PtoP*         Up    29   2d00h    Yes None None
  Area Address:           49.0001
  Neighbor IPv4 Address:  10.3.12.3*
  Adjacency SID:          24006
  Non-FRR Adjacency SID:  24007
  Topology:               IPv4 Unicast
P1             Gi0/0/0/0                *PtoP*         Up    23   2d00h    Yes None None
  Area Address:           49.0001
  Neighbor IPv4 Address:  10.11.12.11*
  Adjacency SID:          24000
  Non-FRR Adjacency SID:  24001
  Topology:               IPv4 Unicast

Total adjacency count: 4
RP/0/RP0/CPU0:P2# 

You see also here Binding-SID of two policies which is dynamically created.

With the command “show mpls forwarding”, you can see that the binding-SIDs are mapped to the corresponding policy.

RP/0/RP0/CPU0:PE1#show mpls forwarding 
Wed Aug 25 16:08:46.229 UTC
Local  Outgoing    Prefix             Outgoing     Next Hop        Bytes       
Label  Label       or ID              Interface                    Switched    
------ ----------- ------------------ ------------ --------------- ------------
16002  16002       SR Pfx (idx 2)     Gi0/0/0/1    10.1.12.12      504         
16003  16003       SR Pfx (idx 3)     Gi0/0/0/1    10.1.12.12      0           
16011  16011       SR Pfx (idx 11)    Gi0/0/0/1    10.1.12.12      0           
16012  Pop         SR Pfx (idx 12)    Gi0/0/0/1    10.1.12.12      0           
24000  Pop         SR Adj (idx 1)     Gi0/0/0/1    10.1.12.12      0           
24001  Pop         SR Adj (idx 3)     Gi0/0/0/1    10.1.12.12      0           
24002  Pop         SR Adj (idx 1)     Gi0/0/0/0    10.1.11.11      0           
24003  Pop         SR Adj (idx 3)     Gi0/0/0/0    10.1.11.11      0           
24004  Pop         No ID              srte_c_10_ep point2point     0           
24005  Pop         No ID              srte_c_20_ep point2point     0           
24006  24003       SR TE: 1 [TE-INT]  Gi0/0/0/1    10.1.12.12      0           
24009  24003       SR TE: 2 [TE-INT]  Gi0/0/0/1    10.1.12.12      0           
RP/0/RP0/CPU0:PE1# 

Also you can check the path of GOLD and SILVER policy with traceroute command in mpls network.

RP/0/RP0/CPU0:PE1#traceroute sr-mpls nil-fec policy binding-sid 24004      
Wed Aug 25 16:10:24.080 UTC

Tracing MPLS Label Switched Path with Nil FEC with Binding SID Label [24004], timeout is 2 seconds

Codes: '!' - success, 'Q' - request not sent, '.' - timeout,
  'L' - labeled output interface, 'B' - unlabeled output interface, 
  'D' - DS Map mismatch, 'F' - no FEC mapping, 'f' - FEC mismatch,
  'M' - malformed request, 'm' - unsupported tlvs, 'N' - no rx label, 
  'P' - no rx intf label prot, 'p' - premature termination of LSP, 
  'R' - transit router, 'I' - unknown upstream index,
  'X' - unknown return code, 'x' - return code 0

Type escape sequence to abort.

  0 10.1.12.1 MRU 1500 [Labels: 24003/explicit-null Exp: 0/0]
L 1 10.1.12.12 MRU 1500 [Labels: implicit-null/explicit-null Exp: 0/0] 19 ms
! 2 10.2.12.2 33 ms

RP/0/RP0/CPU0:PE1#traceroute sr-mpls nil-fec policy binding-sid 24005
Wed Aug 25 16:10:28.633 UTC

Tracing MPLS Label Switched Path with Nil FEC with Binding SID Label [24005], timeout is 2 seconds

Codes: '!' - success, 'Q' - request not sent, '.' - timeout,
  'L' - labeled output interface, 'B' - unlabeled output interface, 
  'D' - DS Map mismatch, 'F' - no FEC mapping, 'f' - FEC mismatch,
  'M' - malformed request, 'm' - unsupported tlvs, 'N' - no rx label, 
  'P' - no rx intf label prot, 'p' - premature termination of LSP, 
  'R' - transit router, 'I' - unknown upstream index,
  'X' - unknown return code, 'x' - return code 0

Type escape sequence to abort.

  0 10.1.12.1 MRU 1500 [Labels: 24003/explicit-null Exp: 0/0]
L 1 10.1.12.12 MRU 1500 [Labels: implicit-null/explicit-null Exp: 0/0] 31 ms
! 2 10.2.12.2 21 ms
P/0/RP0/CPU0:PE1#

With command “show mpls traffic-eng topology”, you can check that TE metric is equal to IGP metric by default.

In the configuration of segment-routing traffic-engineering dynamic path, you can three other options in addition to “igp metric” which they are “te metric”, “hopcount” and also “delay”.

RP/0/RP0/CPU0:PE1(config)#segment-routing traffic-eng 
RP/0/RP0/CPU0:PE1(config-sr-te)#policy GOLD
RP/0/RP0/CPU0:PE1(config-sr-te-policy)#color 10 end-point ipv4 2.2.2.2
RP/0/RP0/CPU0:PE1(config-sr-te-policy)#candidate-paths
RP/0/RP0/CPU0:PE1(config-sr-te-policy-path)#preference 100
RP/0/RP0/CPU0:PE1(config-sr-te-policy-path-pref)#dynamic
RP/0/RP0/CPU0:PE1(config-sr-te-pp-info)#metric
RP/0/RP0/CPU0:PE1(config-sr-te-path-metric)#type ?
  hopcount  Hopcount metric type
  igp       IGP metric type
  latency   Latency metric type
  te        TE metric type
RP/0/RP0/CPU0:PE1(config-sr-te-path-metric)#       type 
Back to: Segment Routing > segment routing traffic engineering

Leave a Reply

Your email address will not be published. Required fields are marked *


Post comment