Seamless MPLS Introduction
- Fatih Kacmaz
- Nov 30, 2021
- 4 min read
Updated: Dec 1, 2021
Seamless MPLS is a mechanism which ensures extending the Core domain and integrating Aggregation and Access domains into single domain. Seamless MPLS uses the ‘divide-and-conquer’ approach, where large portions of the network are divided into many smaller sections. It doesn’t require any new protocol, and uses existing IGP, MPLS, and BGP protocols.
S-MPLS architecture is described in “draft-ietf-mpls-seamless-mpls-07”
At this very first post of mine, I will depict a solution for interconnecting a multi-region network via a core region belonging to the same AS. Depending on the service deployment, the service endpoints might be in the same region, or across different regions. Creating a service at one metro region, and extending it into another metro region, would require “stitching points” at the intermediate points.
Single MPLS provides the following advantages:
No boundaries (Seamless)
Single MPLS Domain
End-to-end Services
Scalable and resilient
A typical S-MPLS network consists of multiple regions: In this case we have a ‘Metro 1’ region, a ‘Core’ region, and a ‘Metro 2’ region. To make this scalable from a transport and service point of view, S-MPLS introduces hierarchy; the transport layer consists of inter-region tunnels and intra-region tunnels.
Figure-1 : End-to-End Tunnel Topology
In the above topology (Figure-1), orange dot represents the service (L2VPN or L3VPN) provisioned on different Metro Regions. Green tunnel depicts the Intra-Area (or AS) tunnel, for which the label is signaled via LDP or RSVP-TE. Purple tunnel, on the other hand, is end-to-end tunnel using the BGP Label.
Inter-region transport tunnels make use of Labeled BGP (described in RFC3107). Labeled BGP transport tunnels provide the PE-to-PE reachability across regions and provide the inner tunnel label of the transport layer hierarchy.
Intra-region transport tunnels use Label Distribution Protocol (LDP) or Resource Reservation Protocol Traffic Engineering (RSVP-TE) as signaling protocol in the MPLS control plane. Intra-region transport tunnels provide the outer tunnel label of the transport label hierarchy.
Seamless MPLS Solution
The following diagram is a sample topology, where 10 Nokia 7750 Service Routers (SR) are connected in a hierarchical fashion. Point-to-point interfaces are also seen below:
Figure-2 : Lab Topology
The System IP Addresses are liste at the below table:
Regarding the sample topology, some design highlights are listed below:
Agr-PE-11/12/13/14 routers are ISIS Level-1-only routers, whereas ABR-1/2/3/4 routers are L1|L2 routers. P1 and P2 routers are Level-2-only routers.
ABR devices export System IP addresses of the Aggregation devices, which are behind their border, to the Core domain.
RSVP-TE is enabled on each point-to-point interface.
Manual LSPs are created from each Aggregation router to its two ABR nodes. Besides, ABR nodes have full-mesh LSP among each other.
P nodes are Route Reflectors of the Core domain, and ABR nodes are Route Reflectors of their Aggregation domain. ABR nodes also apply Next-Hop-Self (NHS) towards their clients.
Above design points are shown in the below figure:
Figure-3 : S-MPLS Design
Configuration Details
The Interface configuration details are provided below:
Figure-4 : Interface Configuration
Agr-PE-11
/configure router
interface "system"
address 10.1.1.11/32
no shutdown
exit
interface "to_ABR_1"
address 192.168.111.0/31
description "to_ABR_1"
ldp-sync-timer 60
port 1/1/2
icmp
no mask-reply
no redirects
exit
no shutdown
exit
interface "to_Agr_PE_12"
address 192.168.112.0/31
description "to_Agr_PE_12"
ldp-sync-timer 60
port 1/1/1
icmp
no mask-reply
no redirects
exit
no shutdown
exit
autonomous-system 65001ISIS configuration details are shown below:
ABR-1
/configure router isis 1
level-capability level-1/2
area-id 49.0000.01
traffic-engineering
reference-bandwidth 1000000
all-l1isis 01:00:5e:01:01:01
all-l2isis 01:00:5e:01:01:02
iid-tlv-enable
loopfree-alternates
exit
level 1
wide-metrics-only
exit
level 2
wide-metrics-only
exit
interface "system"
no shutdown
exit
interface "to_ABR_2"
interface-type point-to-point
no shutdown
exit
interface "to_Agr_PE_11"
interface-type point-to-point
no shutdown
exit
interface "to_P_1"
interface-type point-to-point
no shutdown
exit
no shutdownOn the Aggregation-PE routers, L3VPN services are created. VPRN ID of this service is 1001. There is a loopback interface on each Agr-PE device.
Agr-PE-11
/configure service
vprn 1001 name "1001" customer 1 create
interface "Loopback_1001" create
address 10.11.11.1/32
loopback
exit
bgp-ipvpn
mpls
auto-bind-tunnel
resolution-filter
exit
resolution filter
exit
route-distinguisher 10.1.1.11:1001
vrf-target target:65001:1001
no shutdown
exit
exit
no shutdownBelow configuration is showing the BGP configuration of ABR-1 device:
ABR-1
/configure router bgp
advertise-inactive
enable-peer-tracking
rapid-withdrawal
split-horizon
next-hop-resolution
labeled-routes
transport-tunnel
family label-ipv4
resolution-filter
ldp
rsvp
exit
resolution filter
exit
exit
exit
exit
group "to_P"
family vpn-ipv4 label-ipv4
min-route-advertisement 2
next-hop-self
export "Export_System"
peer-as 65001
advertise-inactive
neighbor 10.0.0.11
description "P-1"
exit
neighbor 10.0.0.22
description "P-2"
exit
exit
group "to_Agr"
family vpn-ipv4 label-ipv4
min-route-advertisement 2
next-hop-self
cluster 10.0.0.1
peer-as 65001
advertise-inactive
neighbor 10.1.1.11
description "Agr-PE-11"
exit
neighbor 10.1.1.12
description "Agr-PE-12"
exit
exit
no shutdown
exit
exitThe show command output below is showing the VRF route table of this VPRN:
A:Agr_PE_11# show router 1001 route-table
===============================================================================
Route Table (Service: 1001)
===============================================================================
Dest Prefix[Flags] Type Proto Age Pref
Next Hop[Interface Name] Metric
-------------------------------------------------------------------------------
10.11.11.1/32 Local Local 01h32m40s 0
Loopback_1001 0
10.21.21.1/32 Remote BGP VPN 01h30m29s 170
10.2.2.21 (tunneled:BGP) 1000
-------------------------------------------------------------------------------
No. of Routes: 2
Flags: n = Number of times nexthop is repeated
B = BGP backup route available
L = LFA nexthop available
S = Sticky ECMP requested
===============================================================================
A:Agr_PE_11#Figure-8 : L3VPN Route Table
The Service-Label and the BGP label can be seen with the following show command output
Figure-9 : Service Label
Fİgure-10 : BGP-Label
Figure-11: RSVP-TE Label
Pinging from ABR-PE-11 VRF instance to the same VRF in ABR-PE-21 is successeful:
Label information is also captured below:
Figure-12 : Packet Capture Results
LDP-to-BGP Stitching
It can happen that small Access Nodes don’t support Labeled BGP (RFC3107). For this reason, a mechanism is used to do redistribution among protocols: LDP to BGP and BGP to LDP or stitching of an LDP FEC to a BGP labeled route and vice versa.
This allows LDP capable PE devices to offer services to PE routers in other areas or domains without the need to support BGP labeled routes.
LDP DoD (Downstream on Demand) is used at locations where Access PE has low scaling as opposed to the LDP DU (Downstream Unsolicited) used in the rest of the region. Aggregate-prefix-match is also used so that a default route can resolve all the received FECs.
Conclusion
The new network requirements for MBH, residential and business services are bringing seamless-mpls architectures where aggregation and core require connectivity
Seamless-MPLS (aka Unified MPLS) is composed of a set of enabling technologies that provide scalability, fast restoration, and single-end-provisoing in a multi-area or inter-as network.
Comments