Download VPN-B

Document related concepts

Zero-configuration networking wikipedia , lookup

Wake-on-LAN wikipedia , lookup

Routing wikipedia , lookup

Multiprotocol Label Switching wikipedia , lookup

Cisco Systems wikipedia , lookup

Transcript
MPLS VPN Configurations
Khalid Raza
CQFE rev17 Russ Davis
© 1999, Cisco Systems, Inc.
1
Agenda
• Introduction to VPNs concepts
• VPN definitions
• Types of VPNs (Overlay/Peer)
• Comparison between Overlay and Peer
model
• Benefits for MPLS VPNs
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
2
Agenda
• Idea behind VRF, RD, RT
• Route propagation in MP-BGP
• Routing between PE-CE
• MPLS Packet Forwarding
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
3
Agenda
• MPLS configuration
VRF
MP-BGP
PE-CE configuration
Advance configuration
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
4
Agenda
• MPLS topologies
• VPN connectivity
• Design considerations
• Deployment strategies
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
5
VPN/MPLS Concepts
• VPN
Concept is to use the service providers shared
resources connecting multiple customer sites
Technologies such as X.25, Frame-relay which use
virtual circuits to establish end-to-end connection
using shared service of the provider infrastructure
This statistical sharing of resources enables the
service provider to offer low cost services to the
end user
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
6
VPN Terminology
• Provider Network (P-Network)
The backbone under control of a Service
Provider
• Customer Network (C-Network)
Network under customer control
• CE router
Customer Edge router. Part of the Cnetwork and interfaces to a PE router
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
7
VPN Terminology
• Site
Set of (sub)networks part of the Cnetwork and co-located
A site is connected to the VPN backbone
through one or more PE/CE links
• PE router
Provider Edge router. Part of the PNetwork and interfaces to CE routers
• P router
CQFE rev14 Russ Davis
Provider (core) router, without knowledge
of VPN
www.Cisco.com
© 1999, Cisco Systems, Inc.
8
VPN Terminology
Provider core
(P) device
CPE (CE)
Device
Provider Edge
(PE) device
Provider Edge
(PE) device
VPN Site
CPE (CE)
Device
VPN Site
Service Provider Network
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
9
Types of VPNs
• VPN services are offered in two major
ways
Overlay Model where the service provider
provides the virtual connections between sites
Peer model where the service provider
participates in the layer routing of the customer
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
10
VPN Overlay Model
• Service provider network is a connection
of point-to-point links
• Routing within the customer network is
transparent to the service provider
network
• Service provider is responsible purely for
data transport between customer sites
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
11
VPN Overlay Model
• Layer 1 implementation (IP, HDLC,
PPP (customer) - provider gives bit
pipes only
• Layer 2 implementation - service
provider responsible for L2 VC via
ATM, Frame-relay
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
12
VPN Overlay Model
Virtual Circuit
Layer-3 Routing Adjacency
CPE (CE)
Device
VPN Site
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
Provider Edge
(PE) device
Provider Edge
(PE) device
Service Provider Network
www.Cisco.com
CPE (CE)
Device
VPN Site
13
VPN Peer Model
• Both provider and customer network use
same network protocol
• CE and PE routers have a routing
adjacency at each site
• All provider routers hold the full routing
information about all customer networks
• Private addresses are not allowed
• May use the virtual router capability
Multiple routing and forwarding tables
based on Customer Networks
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
14
VPN Peer-to-Peer Model
Layer-3 Routing Adjacency
CPE (CE)
Router
VPN Site
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
Provider Edge
(PE) Router
Layer-3 Routing Adjacency
Provider Edge
(PE) Router
Service Provider Network
www.Cisco.com
CPE (CE)
Router
VPN Site
15
VPN Peer Model
• Peer model used two types of
approach
Shared router
Dedicated router
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
16
VPN Peer Model
• Shared router
Where a common router was used, extensive
packet filtering is used on the PE router to
isolate customer
Service provider allocated addresses out of its
space to the customer and managed the packet
filter to ensure same customer reachability, and
isolation between customers.
High maintenance cost associated with packet
filters
Performance impact due to packet filtering
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
17
Peer-to-Peer Model
Shared Router Approach
PE Routing Table
VPN-A routes
VPN-B routes
VPN-C routes
VPN-A
CE
PE
Paris
VPN-B
CE
London
VPN-C
CE
Munich
interface Serial0/1
description ** interface to VPN-A customer
ip address 192.168.61.6 255.255.255.252
ip access-group VPN-A in
ip access-group VPN-A out
!
interface Serial0/2
description ** interface to VPN-B customer
ip address 192.168.61.9 255.255.255.252
ip access-group VPN-B in
ip access-group VPN-B out
!
interface Serial0/3
description ** interface to VPN-C customer
ip address 192.168.62.6 255.255.255.252
ip access-group VPN-C in
ip access-group VPN-C out
Shared router approach with complex filters
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
18
VPN Peer Model
• Dedicated router
Customer isolation is achieved via dedicated
routers connected to customer
POP edge router filter routing updates between
different provider edge routers
Route filtering is achieved via BGP
Communities
Not cost effective
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
19
Peer-to-Peer Model
Dedicated Router Approach
VPN-A
router bgp 111
neighbor 10.13.1.2 remote-as 111
neighbor 10.13.1.2 route-reflector-client
neighbor 10.13.1.2 route-map VPN-A out
!
route-map VPN-A permit 10
match community-list 75
!
ip community-list 75 permit 111:1
CE
Paris
VPN-B
P Router
CE
VPN-A
VPN-A PE
Brussels
VPN-A routes ONLY
VPN-B
CE
VPN-B PE
London
P Routing Table
VPN-A routes (community 111:1)
VPN-B routes (community 111:2)
Dedicated router approach expensive to deploy
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
20
Comparison Between
the Two Models
• Peer Model
• Overlay Model
Easy to implement
Optimal routing
No knowledge of
customer routing
Easy to provision
additional VPNs
through site
provisioning - no
need for link
provisioning
Isolation between
the two network
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
21
Comparison Between
the Two Models
• Peer Model
• Overlay Model
Optimal routing between
sites requires full mesh
Bandwidth provisioning
Virtual circuits have to
be manually configured
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
Customer
convergence is
depended on SP
routing convergence
Lot of routes with
the provider
networks causes
scalability problems
22
Benefits of MPLS VPNs
• Best of both worlds
• PE participates in routing so you can
achieve optimal routing between sites
• PE isolates customer routing information
like dedicated router solution
• Overlapping addresses are permitted
between customers
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
23
Benefits of MPLS VPNs
• PE router is subdivided into virtual routers
• Similar to the dedicated router approach
• Each customer is assigned independent
routing tables
• IOS does this isolation through the
concept of VRF (Virtual Routing and
Forwarding)
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
24
Benefits of MPLS VPNs
VPN Routing Table
VPN-A
CE
Paris
PE
VPN-A
VRF for VPN-A
CE
IGP &/or BGP
London
VRF for VPN-B
VPN-B
CE
Munich
Global Routing Table
Multiple routing & forwarding instances (VRFs) provide
the separation
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
25
Problem
• How to propagate routing across the
network between the PE devices?
• We need a routing protocol that will
transport the customer routes across the
provider network
• Need to maintain the independency of
customers routing and address space
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
26
Easy and Lazy Answer
• Run multiple routing protocols, one each
for customer
• But PE routers will have to run large
number of routing instances
• Poor P router will have to carry all the VPN
routes
• P routers still will run into overlapping
address problem unless you configure all
the vrfs on the PE router
• Does not scale
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
27
Better Solution
• Run a routing protocol that can
exchange the routing updates only
between PE routers
• P router is protected from customer
routes
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
28
But how to do it ?
• Use BGP to pass the routing information
between PE devices
• Use MPLS labels to exchange packets
between next-hops (PE routers)
• Extend BGP to be able to handle
overlapping addresses
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
29
VPN Routing & Forwarding
Instance (VRF)
• PE routers maintain separate routing tables
Global routing table
contains all PE and P routes (perhaps BGP)
populated by the VPN backbone IGP
VRF (VPN routing & forwarding)
routing & forwarding table associated with one or more
directly connected sites (CE routers)
VRF is associated with any type of interface, whether
logical or physical (e.g. sub/virtual/tunnel)
interfaces may share the same VRF if the connected
sites share the same routing information
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
30
VPN Routing & Forwarding
Instance (VRF)
VPN Routing Table
VPN-A
CE
Paris
PE
VPN-A
VRF for VPN-A
CE
IGP &/or BGP
London
VRF for VPN-B
VPN-B
CE
Munich
Global Routing Table
Multiple routing & forwarding instances (VRFs) provide
the separation
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
31
MPLS/VPN Connectivity Model
• Private addressing in multiple VPNs no
longer an issue
provided that members of a VPN do not use the
same address range
VPN A
London
10.2.1.0/24
Address space for
VPN A and B must be
unique
10.3.3.0/24
Munich
10.2.12.0/24
10.4.12.0/24
Milan
VPN B
CQFE rev14 Russ Davis
Paris
© 1999, Cisco Systems, Inc.
Brussels
10.2.1.0/24
www.Cisco.com
Vienna
10.22.12.0/24
VPN C
32
VPN Routing & Forwarding
Instance (VRF)
• VRF can be thought of as a virtual router
with the following structures:
forwarding table based on CEF
a set of interfaces that use the derived forwarding table
rules to control import/export of routes from/into the VPN
routing table
set of routing protocols/peers which inject information into
the VPN routing table (including static routing)
router variables associated with the routing protocol used
to populate the VPN routing table
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
33
VRF Route Population
• VRF is populated locally through PE and CE
routing protocol exchange
RIP Version 2, OSPF, BGP-4 & Static routing
• Separate routing context for each VRF
routing protocol context (BGP-4 & RIP V2)
separate process (OSPF)
C
E
Site-1
PE
EBGP,OSPF, RIPv2,Static
CE
Site-2
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
34
Local VRF Route Population
VPN-A
CE
Paris
VRF for VPN-A
PE
VPN-A
Which routing
protocol context or
process ?
CE
Global
London
VPN-B
VRF for VPN-B
CE
Munich
Local VRF population driven by routing protocol context
or process (OSPF)
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
35
VRF Route Distribution
• PE routers distribute local VPN information
across the MPLS/VPN backbone
through the use of MP-BGP & redistribution from VRF
receiving PE imports routes into attached VRFs
P Router
CE Router
VPN Site
PE
PE
MP-BGP
CE Router
VPN Site
MPLS/VPN Backbone
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
36
Concept of RD
• If customers have overlapping address,
BGP will treat them is single prefix
• Extend the prefix with a 64-bit prefix
(route-distinguisher)
• Now, with 32 bit IP address and 64 bit RD,
the two overlapping IP address are unique
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
37
Concept of RD
• 32 bit IP prefix is the IPv4 address
• With 64 bit RD, it is now extended to
96 bit and is now VPNv4 address
• This address is exchanged only
between the PE routers via BGP
• This is carried in Multi-Protocol BGP
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
38
Concept of RD
VPN-A
CE
PE router converts it into a 96 bit VPNv4 prefix
PE1
MPLS/VPN Backbone
PE2
VPN-B
MP-BGP
CE
VPN-B
Munich
CE
CQFE rev14 Russ Davis
router sends 32 bit IPv4 prefix
© 1999, Cisco Systems, Inc.
www.Cisco.com
BGP Table
Routes from VPN-A
Routes from VPN-B
39
Processing of RD
• RD is propagated between the PE
routers
• RD is removed by the receiving PE
routers
• CE router receives just the IPv4
prefixes
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
40
Usage of RD
• RD is only used to extend the IP prefix
such that overlapping address are unique
• Simple VPN topologies require single RD
per customer
• In some cases multiple RDs may be
required
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
41
Can RD be the VPN Identifier?
• Yes - it could be a VPN identifier
• Complex topologies require another
component for VPN topologies other
than RD, just like communities are
more flexible.
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
42
Concept of RT
• Sites that have to participate in more than
one VPN- RD is not sufficient
• You need another way of deciding the
membership
• RT was introduced to support complex
topologies such that separation and
grouping is easier
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
43
Concept of RT
• RT is extended BGP communities,
attached to VPNv4 address
• Give more flexibility to the VPN
membership
• Any number of RT can be attached to a
route
• Extended communities are 64 bit values
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
44
Concept of RT
• RTs are either exported or imported
• Export route target are attached to the
route the moment it is converted from IPv4
to VPNv4
• Import RT is used to decide the routes that
would be imported into the VPN
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
45
Routing Within MPLS VPN
• Pass IPv4 to the customer routers
• No VPN routes within the MPLS core (P
routers)
• P routers run IGP and global BGP (if
needed)
• Provider Edge router carries connected
VPN routes and Internet routes
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
46
Routing P-router Perspective
• Runs IGP with all the P and PE
routers in the network
• No MPLS VPN routing information
• Very simple view of the network
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
47
Routing PE-router Perspective
• Exchanges IPv4 routes with CE router
• Exchange VPNv4 routes with other PE
routers
• Run common IGP with P router and also
internet BGP with P routers (if needed)
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
48
Routing Table on PE Router
• PE router has to maintain number of
routing tables
• Global routing table (IGP, Internet routes)
• VRF routing information for VPNs
connected
• VRF routing is populated via CE and other
PE routes
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
49
PE to PE Route
Information Flow
• PE router creates VPNv4 update
• Adds extended community attribute (RT,
SOO)
• All other BGP attributes
• Received route is imported into
appropriate VRF according to RT values
• Routes installed into VRF are propagated
to CE routers
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
50
MP-BGP Update
• Any other standard BGP attribute
Local Preference
MED
Next-hop
AS_PATH
Standard Community
• A Label identifying:
The outgoing interface or VRF where a lookup
has to be performed (aggregate/connected)
The BGP label will be the second label in the
label stack of packets travelling in the core
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
51
VRF Population of MP-BGP
ip vrf VPN-A
route-target import VPN-A
VPN-v4 update is translated
into IPv4 address and put
into VRF VPN-A as RT=VPNA and optionally advertised
to CE-2
PE-1
CE-1
Paris
VPN-v4 update:
RD:1:27:149.27.2.0/24,
Next-hop=PE-1
SOO=Paris, RT=VPN-A,
Label=(28)
PE-2
CE-2
London
• Receiving PE routers translate to IPv4
Insert the route into the VRF identified by the RT
attribute (based on PE configuration)
• The label associated to the VPN-V4 address will be
set on packets forwarded toward the destination
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
52
Routing Between PE-CE
• CE does not need any understanding of
MPLS
• CE needs standard IP software
• Currently EBGP, OSPF, RIP, and static
routing is supported
• PE router looks like a standard corporate
backbone to the CE router
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
53
MPLS/VPN Packet Forwarding
In Label
-
FEC
Out Label
197.26.15.1/32
-
In Label
FEC
Out Label
41
197.26.15.1/32
POP
In Label
-
FEC
197.26.15.1/32
197.26.15.1
PE-1
41
PE-2
Use label implicit-null for
destination 197.26.15.1/32
Paris
149.27.2.0/24
Out Label
Use label 41 for destination
197.26.15.1/32
VPN-v4 update:
RD:1:27:149.27.2.0/24,
NH=197.26.15.1
SOO=Paris, RT=VPN-A,
Label=(28)
London
• PE and P routers have BGP next-hop reachability
through the backbone IGP
• Labels are distributed through LDP corresponding
to BGP Next-Hops or RSVP with Traffic Engineering
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
54
MPLS/VPN Packet Forwarding
• Label Stack is used for packet forwarding
Top label indicates BGP Next-Hop (interior label)
Second level label indicates outgoing interface or VRF
(exterior VPN label)
• MPLS nodes forward packets based on top label
any subsequent labels are ignored
• Penultimate Hop Popping procedures used one
hop prior to egress PE router
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
55
Penultimate Hop Popping
In Label
-
FEC
Out Label
197.26.15.1/32
In Label
41
FEC
Out Label
197.26.15.1/32
In Label
POP
-
FEC
Out Label
197.26.15.1/32
41
197.26.15.1
London
Brussels
Use label implicit-null for
destination 197.26.15.1/32
Paris
Use label 41 for destination
197.26.15.1/32
London# show tag-switching tdp binding 197.26.15.1
tib entry: 197.26.15.1/32, rev 10
local binding: tag: imp-null(1)
remote binding: tsr: 172.16.3.1:0, tag: 41
Brussels# show tag-switching tdp binding 197.26.15.1
tib entry: 197.26.15.1/32, rev 10
local binding: tag: 41
remote binding: tsr: 172.16.3.2:0, tag: imp-null(1)
Brussels# show tag-switching forwarding
Local
tag
41
CQFE rev14 Russ Davis
Outgoing
tag or VC
Pop tag
© 1999, Cisco Systems, Inc.
Prefix
or Tunnel Id
197.26.15.1/32
Bytes tag
switched
0
www.Cisco.com
Outgoing
interface
Se0/0/2
Next Hop
point2point
56
MPLS/VPN Packet Forwarding
In Label
FEC
Out Label
-
197.26.15.1/32
41
VPN-A VRF
149.27.2.0/24,
NH=197.26.15.1
Label=(28)
PE-1
41
Paris
149.27.2.0/24
28
149.27.2.27
149.27.2.27
London
• Ingress PE receives normal IP packets
• PE router performs IP Longest Match from
VPN FIB, finds iBGP next-hop and imposes
a stack of labels <IGP, VPN>
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
57
MPLS/VPN Packet Forwarding
In Label
28(V)
VPN-A VRF
149.27.2.0/24,
NH=Paris
FEC
Out Label
In Label
FEC
Out Label
149.27.2.0/24
-
41
197.26.15.1/32
POP
VPN-A VRF
149.27.2.0/24,
NH=197.26.15.1
Label=(28)
PE-1
149.27.2.27
28
149.27.2.27
41
Paris
149.27.2.0/24
28
149.27.2.27
149.27.2.27
London
• Penultimate PE router removes the IGP label
Penultimate Hop Popping procedures (implicit-null label)
• Egress PE router uses the VPN label to select
which VPN/CE to forward the packet to
• VPN label is removed and the packet is routed
toward the VPN site
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
58
MPLS/VPN Configuration and
Implementation
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
59
MPLS Configuration
• VRF: Sites requiring same routing
policies share same VRF
IP routing table
CEF forwarding
Route distinguisher
Route Target (export, import)
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
60
MPLS Configuration
• VRF configuration
Step 1. Create VRF
Step 2. Assign an RD
Step 3. RT export
Step 4. RT import
Step 5. Define an interface to a VRF
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
61
MPLS Configuration
•
VRF configuration
Step 1.
Creating a VRF
ip vrf name
Example ip vrf bootcamp
Where bootcamp is just a name like routemap name
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
62
MPLS Configuration
• VRF configurations
Step 2.
Every VRF needs an associated RD
rd route-distinguisher
Could be AS:X or IP address :X
Example: rd 109:12345
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
63
MPLS Configuration
• VRF configuration
Step 3.
Defining a route target that will be exported
with every route that is send from the VRF
Multiple route-target can be attached to a vrf
route-target export RT
Example: route-target export 109:1234
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
64
MPLS Configuration
• VRF configuration
Step 4.
Define a route-target that will be accepted by
the router to be imported into the VRF
route-target import
Example: route-target import 109:1345
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
65
MPLS Configuration
• VRF configuration
Step 5.
Associate an interface to the VRF; this will
remove the interface from the global routing
process
Existing IP address is removed once the
interface is defined to a VRF; you will have to
re-configure the IP address
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
66
MPLS Configuration
• VRF configuration
Ip vrf GREEN
rd 109:145
route-target export 109:145
route-target import 109:145
interface serial 1/0/1
ip forwarding vrf GREEN
ip address 10.1.1.5 255.255.255.252
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
67
MPLS Configuration
• MP-BGP configuration
BGP process is extended to perform three
functions
Tasks are configured in same BGP process
through address families
1. Maintain and exchange global routing information
(IPv4 routing)
2. VPNv4 routing
3. VRF routing exchange with CE
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
68
MPLS Configuration
• MP-BGP configurations
Global neighbor are configured under the
global BGP process (All P and PE neighbors)
These neighbors need to be activated under
the appropriate address family according to
requirements
VRF specific neighbors are defined under the
corresponding VRFs
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
69
MPLS Configuration
• MP-BGP configurations
Step 1. Configure neighbors and their
parameters under the global process
Step 2. Configure address family VPNv4
Step 3. Activate neighbors to carry VPNv4
routes
Step 4. Activate the VPNv4 specific parameters
under the address family (filter, etc.)
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
70
MPLS Configuration
• MP-BGP configurations
Step 1. Configure BGP process
router bgp 110
neighbor 131.108.1.1 remote-as 110
neighbor 131.108.1.1 update-source loopback 0
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
71
MPLS Configuration
• MP-BGP Configurations
Step 2. Configure the address family, activate the
neighbor under the address family for VNPv4
routes. Neighbor that was defined earlier under
main BGP process
address-family vpnv4
neighbor 131.108.1.1 activate
neighbor 131.108.1.1 next-hop-self
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
72
MPLS Configuration
• Let’s talk a little about the IPv4
address family
Address-family IPv4 is same is your
regular BGP process
Configurations done under this family
will be added to the global BGP
configurations
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
73
MPLS Configuration
no bgp default ipv4 unicast
• Disables the default behavior of IPv4 route
propagation
• Activate the neighbors that need to get
IPv4 routes
• Isolation of VPNv4 and IPv4 routes such
that few neighbors get both and few
receive VPnv4 only
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
74
MPLS Configuration
• Example: 3 neighbors: two of them
need IPv4 routes, one does not
• Requirements
Neighbor 131.108.1.1 (IPv4, VPNv4)
Neighbor 131.108.1.2 (IPv4 only)
Neighbor 131.108.1.3 (VPNv4 only)
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
75
MPLS Configuration
Router bgp 110
No bgp default ipv4 unicast
Neighbor 131.108.1.1 remote-as 110
Neighbor 131.108.1.2 remote-as 110
Neighbor 131.108.1.3 remote-as 110
Neighbor 131.108.1.1 activate
Neighbor 131.108.1.2 activate
Address-family vpnv4
Neighbor 131.108.1.1 activate
Neighbor 131.108.1.3 activate
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
76
MPLS Configuration
• Configuring PE-CE Routing
BGP between PE-CE
RIP between PE-CE
OSPF between PE-CE
Static routes
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
77
MPLS Configuration
• BGP/RIP require single routing process
• Distance/path vector no database
separation needed; done through addressfamilies
• OSPF requires a separate routing process
for each VRF to maintain a separate
database
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
78
MPLS Configuration
• All non-BGP VRF routes have to be
redistributed
• No sync is default
• No auto summary is default
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
79
MPLS Configuration
• BGP
Define the neighbor under the address-family
vrf and not under the global BGP
router bgp 110
!
address-family ipv4 vrf Green
neighbor 10.1.1.1 remote-as 115
neighbor 10.1.1.1 activate
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
80
MPLS Configuration
• RIP
Single routing process
RIP parameters in each VRF
router rip
version 2
address-family ipv4 vrf BLUE
network 10.0.0.0
redistribute bgp 110 metric transparent
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
81
MPLS OSPF
• IGP-BGP redistribution is done by
MPLS
• Not a very good thing for OSPF
• Routes redistributed in OSPF are
external
• Single LSA for every external route
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
82
MPLS OSPF
• If all the routes are carried as
external
• Route summarization would be a
problem
• Stub areas would be hard to
implement
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
83
MPLS OSPF
• MPLS VPNs needed to be extended to
carry OSPF information
• Per se create a concept of super backbone
• Super backbone is created with MP-BGP
between the PE-routers
• This super backbone is between the PE
routers; it is transparent to OSPF
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
84
MPLS OSPF
MPLS BGP
backbone
VPN-A
Area 0
VPN-A
VPN-B
CE
CE
Area 1
Paris
CE
London
VPN-A
VPN-B
CE
Area 2
Area 0
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
85
MPLS OSPF
• OSPF between sites does not use normal
OSPF-BGP redistribution
• Internal OSPF routes are kept internal to
OSPF
• External routes are kept external
• OSPF metrics are preserved
• MPLS OSPF backbone is transparent to
CE OSPF that runs standard software
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
86
MPLS OSPF
• PE routers act as ABRs
• In the case of no stub area, PE
routers also act as ASBRs
• For CE routers’ perspective, send an
inter-area route into the connected
area
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
87
MPLS OSPF
• Intra-area OSPF routes are
redistributed into BGP by the PE
router
• Route Summarization can be done at
the redistribution point by the PE
router
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
88
MPLS OSPF
• Super backbone acts just like area 0 in
regular OSPF
• Redistributed routes at the PE routers
appear as inter-area routes
• Routes from one area 0 site into another
area 0 sites appear as inter-area routes
• Redistributed intra- and inter-area routes
appear as inter-area routes; external still
appear as external
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
89
MPLS OSPF
• For MP-BGP, extended community of
0x8000 is used
• OSPF cost is copied as MED for BGP
• LSA type and metric are carried
across
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
90
MPLS OSPF
• OSPF-BGP loop avoidance
MPLS BGP
backbone
OSPF route
Redistributed into BGP
PE1
VPN-A
Area 0
Paris
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
PE3
VPN-A
VPN-B
PE2
VPN-A
VPN-B
CE
Area 0
www.Cisco.com
91
MPLS OSPF
• PE1 learns the route via OSPF intra-area
• PE1 advertises the route to PE2 and PE3
via MP-BGP
• One of the PE router redistributes it first
(sort of race condition)
• PE2 sends the route to PE3 via OSPF
summary LSA
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
92
MPLS OSPF
• PE3 removes the iBGP route for the
destination and installs the OSPF summary
route, due to lower admin distance
• You can solve the problem by lowering the
administrative distance of iBGP to be
less… not a clean solution
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
93
MPLS OSPF
• To solve this problem a (Down bit) has
been added to option field of the header
like ISIS TLV 135
• PE router sets the down bit when
redistributing routes from MP-BGP to
OSPF
• PE router will never redistribute OSPF
route back into BGP with down bit set
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
94
MPLS OSPF
• Double redistribution loop is still possible
• When the CE does redistribution between
domains and the down bit is lost
• For this purpose, tag field is used as done
by standard BGP-OSPF redistribution
• PE routers never redistributes OSPF
routes with Tag field equal to their own AS
number into MP-BGP
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
95
MPLS Configuration
• OSPF
Configuration is still simple
router ospf 110 vrf RED
network 10.1.0.0 0.0.255.255 area 0
redistribute bgp 110
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
96
MPLS IS-IS
• VPN backbone is treated as a level
above L2
• All L1/L2 routes will be redistributed
into BGP at the PE router
• New extended community in BGP
0x0006
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
97
MPLS IS-IS
• Same as route leaking concept: don’t
send out IS-IS back into BGP if
UP/Down bit is set
• Don’t send route if the route in the
table is not learned via IS-IS
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
98
MPLS IS-IS
• At the receiving site redistribute the
route into IS-IS with UP/Down bit set
• Same concept as separation of
LSDB: one DB can belong to one VPN
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
99
MPLS IS-IS
• Configuration is similar to OSPF
router isis tag1 vrf vpn-blue
net 49.0001.1201.0003.0001.00
redistribute bgp 65000 metric transparent level-1-2
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
100
MPLS Configuration
• Static
Used to configure VRF specific routes
Always need to specify the interface
even though you have the next-hop
ip route vrf YELLOW 10.1.0.0 255.255.0.0 10.1.1.5 serial 2/0
CQFE rev14 Russ Davis
© 1999, Cisco Systems, Inc.
www.Cisco.com
101