PART-04 (Multicast Anycast RP based VXLAN Configuration)

In the case of static ingress replication, the VTEPs need to be manually configured, which creates extra
overhead for the network administrator. As a result, it’s not convenient for large and scalable
infrastructures. Instead of static ingress replication, we can use multicast at the overlay, where broadcast,
unknown unicast, and multicast (BUM) traffic is forwarded to the configured specific multicast group.

Before setting up the lab, it's important to first understand multicast and the need for Anycast RP with
routing protocols. I am sharing a brief explanation of multicast here to clarify why Anycast RP is required
and its significance when configuring a VXLAN overlay with multicast as the data plane.

As we know unicast is the way to send the network traffic on-to-on and for broadcast is way to send the
network traffic one-to-all whereas multicast is the way to send the network traffic one to many but only
them who are part of the multicast group. Multicast is especially efficient for applications where data
needs to be sent to multiple recipients without flooding the entire network, such as live video streaming,
financial market data feeds, and online gaming.
During multimedia traffic transmission using unicast, the server sends a separate copy of the packet to
each client.

However, with broadcast, during multimedia traffic transmission, the server sends a single copy of the
packet to all hosts on the network, including those that are not interested in the multicast traffic.
During multimedia traffic transmission using multicast, the server sends a single copy of the packet to
a special group address, where only hosts interested in receiving the multicast traffic will join.

Brief summarization, advantage and disadvantage of the unicast, Broadcast and multicast for
transmitting the multicast traffic.

Unicast:
In unicast, bandwidth usage increases proportionally with the number of users. It is suitable for a few
clients but not ideal for large or scalable environments.

Broadcast:
With broadcast, only one copy of the packet is sent, but it is delivered to all hosts on the network, even
including those that are not interested in the multicast traffic. As a result, broadcast is not recommended
for data, voice, or video transmission to multiple receivers because even uninterested hosts must process
the packet before discarding it.

Multicast:
Multicast sends packets only to interested hosts. Since a single copy of the packet is sent for all interested
hosts, server resource utilization is optimized. It saves bandwidth and controls network traffic by making
the network replicate packets only when necessary. Routers handle packet replication and delivery to the
receivers.
Example of the multicast traffic with and without multicasting

How to work multicast routing protocol?

• Router to Router transmitting the multicast traffic using multicast routing protocol.
• IGMP uses between the router and host.
Brief introduction of the Multicast routing protocol:

Multicast routing protocols are network protocols designed to manage multicast traffic from its source
to receivers. Unlike unicast routing protocols, multicast routing protocols don’t exchange routing
information among neighbors. Instead, they focus on efficient delivery of data to multiple receivers.

Multicast routing protocols are primarily classified into two categories:

1. Intra-domain Multicast Routing Protocols: These protocols handle multicast traffic within a single
Autonomous System (AS). Types of Intra-Domain Multicast Routing Protocols:
• Distance Vector Multicast Routing Protocol (DVMRP)
• Multicast Open Shortest Path First (MOSPF)
• Protocol Independent Multicast (PIM): PIM has two mode

• PIM Dense Mode (uses source tree)

• Builds a source tree using the IGP shortest path between the sender and receiver.
• Uses a "flood and prune" method to send multicast traffic along the previously built
source tree.
• Suitable for small environments due to its flood-and-prune approach.
• PIM Sparse Mode (uses both source and shared trees)
• Builds a shared tree using the IGP shortest path between the source to the RP
(Rendezvous Point) and from the RP to the receiver.
• Initially, traffic flows over the shared tree, but once the RP is learned, sender and
receiver communicate directly using the shortest path.

2. Inter-domain Multicast Routing Protocols: These protocols handle multicast traffic across different
Autonomous Systems (AS). Types of Inter-Domain Multicast Routing Protocols:
• Multicast Border Gateway Protocol (MBGP)
• Multicast Source Discovery Protocol (MSDP)

To understand how PIM Sparse Mode operates, consider these steps:

Step 1: Discover the PIM neighbors.
Step 2: Identify the root of the tree (RP).
Step 3: The RP learns about the source through PIM register messages.
Step 4: The RP learns about receivers through PIM join messages.
Step 5: Builds a shared tree.
In PIM Sparse Mode, the RP serves as the common point for both the source and receiver. Configuring
the RP can be done in two ways:

1. Static Configuration
• The RP address is manually configured on every router.
• All routers must have the same RP address.
• Failover is not possible in this setup.
2. Dynamic Configuration (two methods)
• Auto-RP (Cisco standard)
• BSR (Bootstrap Router) (open standard)

▪ Both methods support multiple RPs, ensuring high availability and load
balancing.
▪ However, failover is slow, making these options less suitable where low latency
and high performance are critical.

Since both Auto RP and BSR have slow failover convergence, making them less suitable for production
environments. This raises the question: how can we resolve this slow failover convergence? The answer
is to use Anycast RP with an routing protocol.

First try to understand about the anycast. Anycast is a network addressing and routing method in which
the same IP address is assigned to multiple devices (or nodes) across different geographic locations. When
a device needs to reach an anycast address, it will be routed to the nearest (based on routing metrics)
node that has the address. This is commonly used for load balancing, redundancy, and improving response
times.
let’s try to understand, how does Anycast RP resolve the slow failover convergence?

In multicast routing, Anycast RP (Rendezvous Point) is a technique where multiple RPs (typically in
different locations or routers) share the same IP address and act as redundant RPs for multicast sources
and receivers. With Anycast RP, each RP announces the same IP address in the network, but routing
protocols ensure that traffic is directed to the closest RP.

How Anycast RP Failover Works

Failover in Anycast RP is typically achieved through a combination of PIM (Protocol Independent

Multicast) with MSDP (Multicast Source Discovery Protocol), as follows:

1. Same IP, Multiple RPs: Each RP is configured with the same IP address, creating an anycast group
for RPs.
2. Routing to the Nearest RP: When a multicast source sends traffic to an RP, the network routes it
to the nearest RP based on routing metrics.
3. Source Information Sharing via MSDP: To ensure all RPs know about active multicast sources,
MSDP is used. MSDP allows RPs to share multicast source information with each other. This means
if a receiver is using a different RP than the source, it can still learn about the source through the
MSDP messages exchanged between RPs.
4. Automatic Failover:
o If an RP fails, the routing protocol (e.g., OSPF or BGP) automatically redirects traffic to the
next closest available RP with the same anycast IP address.

Example below the different steps when using Anycast RP

 • There are 2 RPs for the same group, they use same IP address
• Receivers on the left have a shorter path towards RP1 so they are connected to the shared tree
(in red) of RP1
• Receiver on the right is closer to RP2 so it's connected to the shared tree (in green) of RP2
• MSDP is running between the 2 RPs for sync the RPs information.

When the source starts generating the stream:

1. First-hop router registers the stream to the closest RP, in this case RP1.
2. RP1 advertises to RP2 via an MSDP SA(Source Active)message that there is an active source 'S' for
the group 'G'.
3. If RP2 has a receiver for group 'G', it joins SPT(Shortest Path Tree)of the source 'S' by sending an
(S,G) join towards the source.
4. RP2 forwards on its shared tree the multicast packet it receives in SA message
5. when last-hop router receives the multicast packet, it joins the SPT (assuming it uses default SPT
threshold)

please note that MSDP SA messages are generated periodically (evrery 1 min) as long as source is
active.
In our example, if RP1 fails, there is only one route left for the RP address pointing to RP2 and so all
routers use RP2 as RP. There is only one shared tree (green) and all multicast operations are fully
restored as soon as IGP has converged.

With Anycast RP, RP failover depends only on IGP convergence. RP convergence = IGP convergence

Benefits of Anycast RP:

• Redundancy: Provides multiple RPs for failover, improving reliability and availability.

• Load Balancing: Spreads traffic load across multiple RPs, which helps in balancing the load for
multicast sources and receivers.

• Improved Latency: Traffic is routed to the nearest RP, which can reduce delay and improve
response time
Workaround summary for the VXLAN configuration using the anycast RP without
MSDP (Why is MSDP not required for this lab? The explanation is given in the lab
configuration details.)

STEP-01: underlay network configuration

1. Topology Design
2. Point to point reachability configuration
3. Loop back interface configuration as VTEP
4. IGP configuration for VTEP reachability

Step -02: overlay network configuration

1. Any cast IP configuration and ensure reachability through IGP

2. Configure Anycast IP as PR and mapping the RP with the Fabric
3. Create VNI and map a VLAN with the VNI
4. create the NVE interface

Step-03: Verification

1. underlay
2. overlay
STEP-01: underlay network configuration

Topology: step-01

The topology is designed in a Leaf and Spine architecture to provide better visibility and understanding
of the VXLAN scenario.
To simplify, steps 2, 3, and 4 are combined and given below.

SPINE-01:

feature ospf
router ospf underlay
router-id 1.1.1.1

interface Ethernet1/1
mtu 9216
ip address 10.10.1.1/30
ip ospf network point-to-point
ip router ospf underlay area 0.0.0.0
no shutdown

interface Ethernet1/2
mtu 9216
ip address 10.10.11.1/30
ip ospf network point-to-point
ip router ospf underlay area 0.0.0.0
no shutdown

interface Ethernet1/3
mtu 9216
ip address 10.10.111.1/30
ip ospf network point-to-point
ip router ospf underlay area 0.0.0.0
no shutdown

interface loopback0
ip address 1.1.1.1/32
ip router ospf underlay area 0.0.0.0

SPINE-02:

feature ospf
router ospf underlay
router-id 2.2.2.2

interface Ethernet1/1
mtu 9216
ip address 10.10.2.1/30
ip ospf network point-to-point
ip router ospf underlay area 0.0.0.0
no shutdown
interface Ethernet1/2
mtu 9216
ip address 10.10.22.1/30
ip ospf network point-to-point
ip router ospf underlay area 0.0.0.0
no shutdown

interface Ethernet1/3
mtu 9216
ip address 10.10.222.1/30
ip ospf network point-to-point
ip router ospf underlay area 0.0.0.0
no shutdown

interface loopback0
ip address 2.2.2.2/32
ip router ospf underlay area 0.0.0.0

LEAF-01:

feature ospf
router ospf underlay
router-id 11.11.11.11

interface Ethernet1/1
mtu 9216
ip address 10.10.1.2/30
ip ospf network point-to-point
ip router ospf underlay area 0.0.0.0
no shutdown

interface Ethernet1/2
mtu 9216
ip address 10.10.2.2/30
ip ospf network point-to-point
ip router ospf underlay area 0.0.0.0
no shutdown

interface loopback0
ip address 11.11.11.11/32
ip router ospf underlay area 0.0.0.0
LEAF-02:

feature ospf
router ospf underlay
router-id 22.22.22.22

interface Ethernet1/1
mtu 9216
ip address 10.10.11.2/30
ip ospf network point-to-point
ip router ospf underlay area 0.0.0.0
no shutdown

interface Ethernet1/2
mtu 9216
ip address 10.10.22.2/30
ip ospf network point-to-point
ip router ospf underlay area 0.0.0.0
no shutdown

interface loopback0
ip address 22.22.22.22/32
ip router ospf underlay area 0.0.0.0

LEAF-03:
feature ospf
router ospf underlay
router-id 33.33.33.33

interface Ethernet1/1
mtu 9216
ip address 10.10.111.2/30
ip ospf network point-to-point
ip router ospf underlay area 0.0.0.0
no shutdown

interface Ethernet1/2
mtu 9216
ip address 10.10.222.2/30
ip ospf network point-to-point
ip router ospf underlay area 0.0.0.0
no shutdown

interface loopback0
ip address 33.33.33.33/32
ip router ospf underlay area 0.0.0.0
Step -02: overlay network configuration

Any cast IP configuration and ensure reachability through IGP (Only for SPINE-01 & 02)

interface loopback12
ip address 12.12.12.12/32
ip router ospf underlay area 0.0.0.0

Configure Anycast IP as PR and mapping the RP with the Fabric

Enable the feature PIM

feature pim

Configure anycast IP as RP (SPINE-01 & 02)

ip pim anycast-rp 12.12.12.12 1.1.1.1

ip pim anycast-rp 12.12.12.12 2.2.2.2

Mapping the RP (SPINE-01 & 02 and LEAF-01, 02 & 03)

ip pim rp-address 12.12.12.12

Enable the PIM all ports that connected with SPINE and LEAF

Interface 1/X
Ip pim sparse-mode

As per our topology, each leaf switch has two uplinks, one to each spine switch. Some leaf switches
connect to one spine switch, while others connect to the second spine switch. This setup is not random;
a hashing algorithm is used to decide which uplink to use. ECMP multicast uses an S-G hash based on the
source and group address. This hashing algorithm is predictable because it doesn’t rely on random values.

To synchronize information between RPs, we can use one of two protocols:

• Multicast Source Discovery Protocol (MSDP)
• PIM
Cisco Nexus switches can use Anycast RP without MSDP. PIM messages are exchanged between spine
switches, allowing all spine switches to be aware of multicast sources. As a result, MSDP is not required
for Cisco Nexus switches.

Create VNI and map a VLAN with the VNI: Step-01 (Only for LEAF-01,02 & 03)

feature vn-segment-vlan-based

vlan 10
vn-segment 101010
create the NVE interface: Step-02

LEAF-01:

feature nv overlay

interface nve1
no shutdown
source-interface loopback0
member vni 101010
mcast-group 239.10.10.10

LEAF-02:

feature nv overlay

interface nve1
no shutdown
source-interface loopback0
member vni 101010
mcast-group 239.10.10.10

LEAF-03:

feature nv overlay

interface nve1
no shutdown
source-interface loopback0
member vni 101010
mcast-group 239.10.10.10

Step-03: Verification

Underlay verification

SPINE-01# show ip ospf neighbors (To check the underlay IGP neighbors relationship)
SPINE-01# show ip ospf route (To check the VTEP route)

LEAF-01# ping REMOTE_VTEP_IP (To check the VTEP reachability)

overlay verification

LEAF-01 # show ip route ospf-underlay (To check the anycast IP reachability)

LEAF-01# show vxlan (To check the vlan mapping with vni id)

LEAF-01#show nve peers (To check the nve tunneling )

LEAF-01#show nve vni (To check the data plane learning status)

LEAF-01#show interface nve 1 (To check the NVE interface)

LEAF-01#show nve interface nve 1 detail (To check the NVE interface details)
LEAF-01# show mac address-table dynamic (To check the flood and learn mechanism on the data plane)

LEAF-01# ping IP_address (Host Reachability check)

Limitation of the VXLAN configuration using Multicast

static ingress replication has issue need to manually configure all VTEPs so there is administrative
overhead. Even it also used Broadcast which generate huge unwanted traffic. But configuration using
the Multicast as data plane it reduces the manual configuration. But still remain the issue of broadcast
which is not expected for the any production environment where performance is crucial.
Configuration

SPINE-01

feature ospf
feature pim
feature vn-segment-vlan-based
feature nv overlay

ip pim rp-address 12.12.12.12

ip pim anycast-rp 12.12.12.12 1.1.1.1

ip pim anycast-rp 12.12.12.12 2.2.2.2

interface Ethernet1/1
mtu 9216
ip address 10.10.1.1/30
ip ospf network point-to-point
ip router ospf underlay area 0.0.0.0
ip pim sparse-mode
no shutdown

interface Ethernet1/2
mtu 9216
ip address 10.10.11.1/30
ip ospf network point-to-point
ip router ospf underlay area 0.0.0.0
ip pim sparse-mode
no shutdown

interface Ethernet1/3
mtu 9216
ip address 10.10.111.1/30
ip ospf network point-to-point
ip router ospf underlay area 0.0.0.0
ip pim sparse-mode
no shutdown

interface loopback0
ip address 1.1.1.1/32
ip router ospf underlay area 0.0.0.0
ip pim sparse-mode

interface loopback12
ip address 12.12.12.12/32
ip router ospf underlay area 0.0.0.0
 ip pim sparse-mode

router ospf underlay

router-id 1.1.1.1

SPINE-02

feature ospf
feature pim
feature vn-segment-vlan-based
feature nv overlay

ip pim rp-address 12.12.12.12

ip pim anycast-rp 12.12.12.12 1.1.1.1

ip pim anycast-rp 12.12.12.12 2.2.2.2

interface nve1
source-interface loopback0
member vni 101010
ingress-replication protocol static
peer-ip 11.11.11.11
peer-ip 22.22.22.22
peer-ip 33.33.33.33

interface Ethernet1/1
mtu 9216
ip address 10.10.2.1/30
ip ospf network point-to-point
ip router ospf underlay area 0.0.0.0
ip pim sparse-mode
no shutdown

interface Ethernet1/2
mtu 9216
ip address 10.10.22.1/30
ip ospf network point-to-point
ip router ospf underlay area 0.0.0.0
ip pim sparse-mode
no shutdown

interface Ethernet1/3
mtu 9216
ip address 10.10.222.1/30
ip ospf network point-to-point
ip router ospf underlay area 0.0.0.0
ip pim sparse-mode
no shutdown
interface loopback0
ip address 2.2.2.2/32
ip pim sparse-mode

interface loopback12
ip address 12.12.12.12/32
ip router ospf underlay area 0.0.0.0
ip pim sparse-mode

router ospf underlay

router-id 2.2.2.2

LEAF-01

feature ospf
feature pim
feature vn-segment-vlan-based
feature nv overlay

ip pim rp-address 12.12.12.12

vlan 1,10
vlan 10
vn-segment 101010

interface nve1
no shutdown
source-interface loopback0
member vni 101010
mcast-group 239.10.10.10

interface Ethernet1/1
mtu 9216
ip address 10.10.1.2/30
ip ospf network point-to-point
ip router ospf underlay area 0.0.0.0
ip pim sparse-mode
no shutdown

interface Ethernet1/2
mtu 9216
ip address 10.10.2.2/30
ip ospf network point-to-point
ip router ospf underlay area 0.0.0.0
ip pim sparse-mode
no shutdown

interface Ethernet1/3
 switchport
switchport access vlan 10
no shutdown

interface loopback0
ip address 11.11.11.11/32
ip router ospf underlay area 0.0.0.0
ip pim sparse-mode

router ospf underlay

router-id 11.11.11.11

LEAF-02

feature ospf
feature pim
feature vn-segment-vlan-based
feature nv overlay

ip pim rp-address 12.12.12.12

vlan 1,10
vlan 10
vn-segment 101010

interface nve1
no shutdown
source-interface loopback0
member vni 101010
mcast-group 239.10.10.10

interface Ethernet1/1
mtu 9216
ip address 10.10.11.2/30
ip ospf network point-to-point
ip router ospf underlay area 0.0.0.0
ip pim sparse-mode
no shutdown

interface Ethernet1/2
mtu 9216
ip address 10.10.22.2/30
ip ospf network point-to-point
ip router ospf underlay area 0.0.0.0
ip pim sparse-mode
no shutdown

interface Ethernet1/3
 switchport
switchport access vlan 10
no shutdown

interface loopback0
ip address 22.22.22.22/32
ip router ospf underlay area 0.0.0.0
ip pim sparse-mode

router ospf underlay

router-id 22.22.22.22

LEAF-03

feature ospf
feature pim
feature vn-segment-vlan-based
feature nv overlay

ip pim rp-address 12.12.12.12

vlan 1,10
vlan 10
vn-segment 101010

interface nve1
no shutdown
source-interface loopback0
member vni 101010
mcast-group 239.10.10.10

interface Ethernet1/1
mtu 9216
ip address 10.10.111.2/30
ip ospf network point-to-point
ip router ospf underlay area 0.0.0.0
ip pim sparse-mode
no shutdown

interface Ethernet1/2
mtu 9216
ip address 10.10.222.2/30
ip ospf network point-to-point
ip router ospf underlay area 0.0.0.0
ip pim sparse-mode
no shutdown

interface Ethernet1/3
 switchport
switchport access vlan 10
no shutdown

interface loopback0
ip address 33.33.33.33/32
ip router ospf underlay area 0.0.0.0
ip pim sparse-mode

router ospf underlay

router-id 33.33.33.33