H3C S12500AI Switch Series DDC Best Practices

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Contents

About the DDC solution· 1

Background· 1

Weaknesses of classic switches· 1

DDC technology· 1

DDC physical architecture· 2

DDC physical devices· 2

DDC requirements for device connections· 3

DDC data forwarding mechanism·· 4

Data forwarding mechanism·· 4

Cell-based forwarding network· 4

Control plane· 5

Forwarding procedures· 5

DDC configuration examples· 6

Configuring the DDC control plane· 6

Networking configuration· 6

Analysis· 6

Procedures· 7

Configuring the DDC service plane· 9

Configuring an 800G NCP· 9

Configuring an 400G NCP· 10

Configuring tenant isolation in the DDC system·· 12

Configuring ACL-based isolation· 12

Configuring VPN-based isolation· 13

 

About the DDC solution

Background

Weaknesses of classic switches

With the rapid development of big data, cloud computing, and artificial intelligence (AI) technologies, data centers need to accommodate traffic surge, which put immense pressure on their core switches.

Traditionally, modular switches act as core switches in data centers. Each modular switch is an enclosed, centralized, and large chassis. All switch components, such as MPUs, switching modules, and interface modules are housed within a large-size chassis. Although this design allows centralized management, it has limitations in scalability, flexibility, and cost-effectiveness. As switching chips advance and switching capacity increases (from 100 G to 400 G), traditional modular switches have a significant increase in power consumption. For example, a 16-slot modular switch with all 400-G ports requires a power supply of up to 40000 to 50000 watts. This requirement makes upgrading devices in old data centers a significant challenge, especially when the power supply cannot meet this requirement.

DDC technology

Diversified Dynamic-Connectivity (DDC) is an innovative network architecture that breaks away from the classic switch design (centralized and modular switch). It adopts a distributed, decoupled approach to enhance flexibility and scalability in data center networks. This technology provides breakthroughs both in physical appearance and data forwarding.

·     Physical appearance: DDC decomposes a large modular switch into multiple fixed-port switches.

As shown in Figure 1, DDC splits a classic modular switch into smaller, independent components, fixed-port switches, which allows for distributed deployment of network functions. This process is similar as decomposing a “big iron lump” into “building blocks”, enabling distributed deployment of network functions. Those fixed-port switches act as forwarding interface modules or switching fabric modules and are deployed across multiple cabinets. This approach not only provides better cooling management and power management consumption control, but also allows flexible and convenient network deployment by eliminating limitations in device upgrade and space expansion.

Figure 1 Classic switch vs DDC device

 

·     Data forwarding: DDC integrates the physical connections among multiple fixed-port switches into a cell-based forwarding network. In such a network, data packets are forwarded as quickly and efficiently as they are forwarded within a modular switch.

DDC physical architecture

DDC physical devices

Figure 2 DDC physical architecture

 

As shown in Figure 2, H3C's DDC solution decouples the classic design of modular switch into two types of physical devices including network connectivity fabric (NCF) and network connectivity processor (NCP).

NCF functions similarly as the switching fabric module of a modular switch, which is used for transparent packet transmission. NCP functions similarly as the service module and MPU of a modular switch, which is responsible for processing protocol packets and forwarding service packets. Data packets are transmitted between NCPs and NCFs through SFIs.

Table 1 shows the port matrix for the S12500AI switch series.

Table 1 Port matrix for the S12500AI switch series

Product series	Chassis type	Product model	Available interfaces
H3C S12500AI switch series	NCF	S12500AI-NCFN	·     128 x OSFP800 SFIs
	NCP	S12500AI-36DH20EP-NCPN	·     36 x QSFP 112 service interfaces ·     20 x OSFP800 SFIs
		S12500AI-18EP20EP-NCPN	·     18 x OSFP800 service interfaces ·     20 x OSFP800 SFIs

 

NOTE: An NCF supports four interface modules, with each module providing 32 SFs. Therefore, an NCF can provide a total of 128 SFIs.

 

DDC requirements for device connections

DDC has the following device connection requirements:

·     Each NCP is connected directly to the NCFs. No additional connections are required between NCPs or between NCFs.

·     The number of NCPs and NCFs depends on the network size. As a best practice, make sure the ratio of NCPs to NCFs is 6:1.

·     The maximum cable length between NCP and NCF is 300 meters. As a best practice, make sure the cable lengths between NCP and NCF are the same (the difference in cable length between NCP and NCF should not exceed 50%).

Figure 3 DDC network topology

 

DDC data forwarding mechanism

Data forwarding mechanism

In classic Ethernet networks, traffic forwarding is implemented with a combination of routing protocols and ECMP, which achieves load balancing to some extent. However, factors such as packet count, traffic size, N-tuple, and hash deviation might affect hash results, leading to imbalanced load distribution—especially during forwarding of large flows (elephant flows). Additionally, traditional ECMP relies on hop-by-hop routing and cannot dynamically sense the available bandwidth of each forwarding path, which might cause packet loss during traffic congestion.

DDC adopts cell switching to achieve balanced traffic forwarding across NCPs and improve forwarding efficiency. This solution segments packets into cells of the same size and distributes them across multiple paths. Meanwhile, DDC also supports VOQ-based path awareness. Before forwarding data packets, NCPs determine whether the path bandwidth is available. They perform load balancing only across usable paths to achieve non-blocking switching.

Cell-based forwarding network

Based on SFIs and internal protocol interactions, the interconnected NCPs and NCFs in a DDC system automatically form a cell-based forwarding network. In this network, traffic is forwarded through cell-based forwarding. For devices outside the DDC system, those NCPs and NCFs function as a single device at the forwarding plane.

To manage the entire cell network, each NCP or NCF is configured with a unique member ID. Meanwhile, each NCP service interface is assigned a Systemport ID, which uniquely identifies that service interface within the entire cell-based forwarding network. Data forwarding is implemented based on Systemport IDs.

Control plane

Each NCP is equipped with a CPU virtual interface (OSF interface) facing the cell network. These virtual interfaces are logically connected to a Layer 2 network and reside in the same broadcast domain.

Additionally, all NCP CPU interfaces should run on the same IP address range to establish BGP peer relationships. This enables synchronization of entries such as ARP entries and routes, ensuring effective communication and coordination among NCPs.

Once the control plane is established, NCPs build forwarding entries. For local forwarding, those entries are associated only with local output interfaces’ IDs. For cross-NCP forwarding, those entries are associated with the Systemport IDs of remote output interfaces and remote encapsulation indexes, which ensures that packets are encapsulated and sent to remote NCPs correctly.

Forwarding procedures

Figure 4 DDC data plane

 

Data is forwarded on the DDC data plane as follows: (In this example, packets are forwarded from Server1 to Server4):

1.     When Server1 sends a packet to Server4, NCP1 first looks up the related forwarding entry to obtain the Systemport and remote encapsulation index.

2.     Based on the SystemPort, NCP1 adds the packet to the related VOQ. If available credits exist, the packet is segmented and encapsulated as cells, and then sent to the upstream NCF through the related SFI.

3.     Upon receiving those cells, the NCF looks up the related local cell forwarding entry based on the cell forwarding information carried in those cells, determines the output interface, and then forwards the cells to NCP4.

4.     After NCP4 receives the cells, it reassembles the packet based on the received cell forwarding information, obtains encapsulation information from the cell header, encapsulates the reassembled packet, and then sends it to Server4 through the designated port.

DDC configuration examples

Configuring the DDC control plane

Networking configuration

As shown in Figure 5, an RoCE network is deployed on the cell-based forwarding network formed by NCPs and NCFs deploys a RoCE network. Each server NIC is directly connected to the related upstream device through a single port. To achieve lossless packet transmission, RDMA application packets must be transmitted by using Queue 2. (Queue 2 is used as an example. The real queue depends on your live network.)

Figure 5 DDC network diagram

 

Analysis

·     Place all NCP OSF ports on the same subnet.

·     Make sure all SFIs used for NCP-NCF-NCP connections are in UP state.

·     Establish L2VPN EVPN peer relationships among NCPs by using OSF port IP addresses.

Procedures

NOTE: By default, all member IDs are 0. Each member device can start, but cannot participate in data forwarding. To change the member ID for a device, use the cloud-cluster member X renumber Y command. The member ID range varies by DDC role as follows: ·     NCP: 0 to 255. ·     NCF: 0 to 40. To avoid conflicts, make sure each node has a unique member ID among nodes with the same DDC role (NCP or NCF). No manual saving is required after configuration changes. The new member ID takes effect automatically after a device reboot.

 

Configure NCP1

# Change the member ID for NCP1.

<NCP1> system-view

[NCP1] cloud-cluster member 0 renumber 1

This command will take effect after the cloud cluster configuration is activated. The command might result in configuration change or loss when it takes effect. Continue? [Y/N]: y

# Reboot the device to have the new member ID take effect.

[NCP1] quit

<NCP1> reboot

Prompt omitted...

Configure NCP2

# Change the member ID for NCP2.

<NCP2> system-view

[NCP2] cloud-cluster member 0 renumber 2

This command will take effect after the cloud cluster configuration is activated. The command might result in configuration change or loss when it takes effect. Continue? [Y/N]: y

# Reboot the device to have the new member ID take effect.

[NCP2] quit

<NCP2> reboot

Prompt omitted...

Configure NCF1

# Change the member ID to 3 for NCF1.

<NCF1> system-view

[NCF1] cloud-cluster member 0 renumber 3

This command will take effect after the cloud cluster configuration is activated. The command might result in configuration change or loss when it takes effect. Continue? [Y/N]: y

# Reboot the device to have the new member ID take effect.

[NCF1] quit

<NCF1> reboot

Prompt omitted...

Configure NCF2

# Change the member ID to 4 for NCF2.

<NCF2> system-view

[NCF2] cloud-cluster member 0 renumber 4

This command will take effect after the cloud cluster configuration is activated. The command might result in configuration change or loss when it takes effect. Continue? [Y/N]: y

# Reboot the device to have the new member ID take effect.

[NCF2] quit

<NCF2> reboot

Prompt omitted...

Detailed configurations at the DDC control plane

NCP1	NCP2	Description
OSF interface settings:
interface OSF0/0/0	interface OSF0/0/0	Enter OSF interface view.
ip address 100.0.0.2 255.255.255.0	ip address 100.0.0.3 255.255.255.0	Assign an IP address for establishing BGP EVPN peer relationships with other NCPs.
Global settings:
ip public-instance	ip public-instance	Create the public instance and enter its view.
route-distinguisher 1:1	route-distinguisher 2:1	Specify a route distinguisher.
vpn-target 2:1 import-extcommunity	vpn-target 2:1 import-extcommunity	Specify an import route target.
vpn-target 2:1 export-extcommunity	vpn-target 2:1 export-extcommunity	Specify an export route target.
address-family ipv4	address-family ipv4	Enable cell encapsulation for subnet routes.
evpn osf routing-enable	evpn osf routing-enable
address-family evpn	address-family evpn	Enable cell encapsulation for ARP routes.
evpn osf routing-enable	evpn osf routing-enable
BGP settings:
bgp 100	bgp 100	Create a BGP instance and enter its view.
router-id 100.0.0.2	router-id 100.0.0.3	Specify a router ID.
peer 100.0.0.3 as-number 100	peer 100.0.0.2 as-number 100	Specify a peer AS number for peering.
address-family ipv4 unicast	address-family ipv4 unicast	NCPs establish L2VPN EVPN peer relationships through their OSF interface IP addresses. The related service interfaces should run on the public network. You can redistribute direct routes and static routes, and establish BGP peer relationships as needed.
import-route direct	import-route direct	Redistribute direct routes.
address-family l2vpn evpn	address-family l2vpn evpn	Create the BGP EVPN address family and enter its view.
peer 100.0.0.3 enable	peer 100.0.0.2 enable	Establish a peer relationship with a peer device.
peer 100.0.0.3 advertise encap-type osf	peer 100.0.0.2 advertise encap-type osf	This configuration is used in conjunction with the cell configuration in the public instance or related VPN instance, which indicates that the device advertises routes with cell encapsulation.
Interface settings:
interface FourHundredGigE1/0/8	interface FourHundredGigE2/0/8	Associate the service interface with the public network.
ip address 145.0.0.1 255.255.255.0	ip address 146.0.0.1 255.255.255.0	Assign an IP address.

 

Configuring the DDC service plane

Configuring an 800G NCP

Analysis

For lossless transmission of RDMA application packets, configure PFC and ECN:

·     PFC:

¡     PFC performs flow control based on queues with different priorities. RDMA packets carry an 802.1p priority value of 5. Therefore, you must enable PFC for 802.1p priority 5.

¡     PFC must be enabled on all NCP ports along the forwarding path for RDMA packets.

·     ECN:

¡     ECN provides end-to-end congestion control. After congestion is detected on the device, the ECN field is marked for related packets. When the receiver receives packets with the ECN field marked, it sends a congestion notification to the sender, notifying the sender to decrease the packet sending rate.

¡     When you configure static ECN, you must enable ECN on all server-facing ports of the NCP.

¡     For ECN to react earlier than PFC, make sure the high-limit value specified in the queue queue-id [ drop-level drop-level ] low-limit low-limit high-limit high-limit [ discard-probability discard-prob ] command is smaller than the PFC backpressure frame triggering threshold.

Restrictions and guidelines

·     Each NCP must be enabled with PFC and ECN.

·     An 800G NCP refers to an NCP with 800G service ports. The device model is S12500AI-18EP20EP-NCPN.

Procedures

NCP1	NCP2	Description
Global settings:
qos wred queue 2 drop-level {0/1/2} low-limit 250000 high-limit 1200000 discard-probability 20	qos wred queue 2 drop-level {0/1/2} low-limit 250000 high-limit 1200000 discard-probability 20	Configure WRED parameters for queue 2 as follows: Specify a drop level of 0/1/2, set the lower limit for the average queue length to 250000, the upper limit for the average queue length to 1200000, and the denominator for drop probability calculation to 20%.
qos wred queue 2 weighting-constant 0	qos wred queue 2 weighting-constant 0	Set the exponent for WRED to calculate the average queue size
qos wred queue 2 ecn	qos wred queue 2 ecn	Enable ECN for queue 2.
Interface settings:
qos trust dscp	qos trust dscp	Set the priority trust mode to DSCP.
priority-flow-control enable	priority-flow-control enable	Enable PFC.
priority-flow-control no-drop dot1p 5	priority-flow-control no-drop dot1p 5	Enable PFC for 802.1p priority 5.
priority-flow-control dot1p 5 headroom 600000	priority-flow-control dot1p 5 headroom 600000	Configure the maximum number of cell resources that can be used in the headroom storage space for queues. As a best practice, set the maximum number to 600000 for 800G NCPs.
priority-flow-control dot1p 5 reserved-buffer 256	priority-flow-control dot1p 5 reserved-buffer 256	Set the PFC reserved threshold.
priority-flow-control no-drop dot1p 5 pause-threshold ratio 12	priority-flow-control no-drop dot1p 5 pause-threshold ratio 12	Configure flow control over traffic with an 802.1p priority of 5 to ensure lossless transmission against network congestion.
priority-flow-control no-drop dot1p 5  pause-threshold-offset 1024	priority-flow-control no-drop dot1p 5 pause-threshold-offset 1024	Set the offset to 1024 between the back pressure frame stopping threshold and triggering threshold.
qos wfq byte-count	qos wfq byte-count	Enable byte-count WFQ.
qos wfq af2 group 1 byte-count 60	qos wfq af2 group 1 byte-count 60	Add the related queue to a WFQ group and assign a scheduling weight to that queue.
qos wfq af3 group sp	qos wfq af3 group sp	Assign the related queue to the SP group.
qos gts queue 3 cir 400000000 cbs 16000000	qos gts queue 3 cir 400000000 cbs 16000000	Set the CIR to 400000000 kbps for the related queue.

 

Configuring an 400G NCP

Analysis

For lossless transmission of RDMA application packets, configure PFC and ECN:

·     PFC:

¡     PFC performs flow control based on queues with different priorities. RDMA packets carry an 802.1p priority value of 5. Therefore, you must enable PFC for 802.1p priority 5.

¡     PFC must be enabled on all ports along the forwarding path for RDMA packets.

·     ECN:

¡     ECN provides end-to-end congestion control. After congestion is detected on the device, the ECN field is marked for related packets. When the receiver receives packets with the ECN field marked, it sends a congestion notification to the sender, notifying the sender to decrease the packet sending rate.

¡     When you configure static ECN, you must enable ECN on all server-facing ports of the NCP.

¡     For ECN to react earlier than PFC, make sure the high-limit value specified in the queue queue-id [ drop-level drop-level ] low-limit low-limit high-limit high-limit [ discard-probability discard-prob ] command is smaller than the PFC backpressure frame triggering threshold.

Restrictions and guidelines

·     Each NCP must be enabled with PFC and ECN.

·     An 400G NCP refers to an NCP with 400G service ports. The device model is S12500AI-36DH20EP-NCPN.

Procedures

NCP1	NCP2	Description
Global settings:
qos wred queue 2 drop-level {0/1/2} low-limit 250000 high-limit 1200000 discard-probability 20	qos wred queue 2 drop-level {0/1/2} low-limit 250000 high-limit 1200000 discard-probability 20	Configure WRED parameters for queue 2 as follows: Specify a drop level of 0/1/2, set the lower limit for the average queue length to 250000, the upper limit for the average queue length to 1200000, and the denominator for drop probability calculation to 20%.
qos wred queue 2 weighting-constant 0	qos wred queue 2 weighting-constant 0	Set the exponent for WRED to calculate the average queue size
qos wred queue 2 ecn	qos wred queue 2 ecn	Enable ECN for queue 5.
Interface settings:
qos trust dscp	qos trust dscp	Set the priority trust mode to DSCP.
priority-flow-control enable	priority-flow-control enable	Enable PFC.
priority-flow-control no-drop dot1p 5	priority-flow-control no-drop dot1p 5	Enable PFC for 802.1p priority 5.
priority-flow-control dot1p 5 headroom 300000	priority-flow-control dot1p 5 headroom 300000	Configure the maximum number of cell resources that can be used in the headroom storage space for queues. As a best practice, set the maximum number to 300000 for 400G NCPs.
priority-flow-control dot1p 5 reserved-buffer 256	priority-flow-control dot1p 5 reserved-buffer 256	Set the PFC reserved threshold.
priority-flow-control no-drop dot1p 5 pause-threshold ratio 12	priority-flow-control no-drop dot1p 5 pause-threshold ratio 12	Configure flow control over traffic with an 802.1p priority of 5 to ensure lossless transmission against network congestion.
priority-flow-control no-drop dot1p 5  pause-threshold-offset 1024	priority-flow-control no-drop dot1p 5  pause-threshold-offset 1024	Set the offset to 1024 between the back pressure frame stopping threshold and triggering threshold.
qos wfq byte-count	qos wfq byte-count	Enable byte-count WFQ.
qos wfq af2 group 1 byte-count 60	qos wfq af2 group 1 byte-count 60	Add the related queue to a WFQ group and assign a scheduling weight to that queue.
qos wfq af3 group sp	qos wfq af3 group sp	Assign the related queue to the SP group.
qos gts queue 3 cir 200000000 cbs 16000000	qos gts queue 3 cir 200000000 cbs 16000000	Set the CIR to 200000000 kbps for the related queue.

 

Configuring tenant isolation in the DDC system

NOTE: The configuration procedure for 400G devices is the same as that for 800G devices.

 

Configuring ACL-based isolation

Networking configuration

As shown in Figure 6, you can configure ACLs to meet the following isolation requirements:

·     Server1, Server2, and Server3 are isolated from each other and cannot communicate with each other.

·     Server4, Server5, and Server6 are isolated from each other and cannot communicate with each other.

·     Server1 and Server4 can communicate with each other.

·     Server2 and Server5 can communicate with each other.

The ACLs are as follows:

·     On Port1 of NCP1, configure an ACL to allow traffic from Server1 to Server4.

·     On Port2 of NCP1, configure an ACL to allow traffic from Server2 to Server5.

·     On Port1 of NCP2, configure an ACL to allow traffic from Server4 to Server1.

·     On Port2 of NCP2, configure an ACL to allow traffic from Server5 to Server2.

 

NOTE: Port1 and Port2 of NCP1 represent FourHundredGigE1/0/1 and FourHundredGigE1/0/2, respectively. Port1 and Port2 of NCP2 represent FourHundredGigE2/0/1 and FourHundredGigE2/0/2, respectively.

 

Figure 6 Network diagram

 

Procedures

NCP1	NCP2	Description
Global settings:
acl number 3000	acl number 3000	Create an ACL.
rule 5 permit ip source 145.1.1.0 0.0.0.255 destination 146.1.1.0 0.0.0.255	rule 5 permit ip source 146.1.1.0 0.0.0.255 destination 145.1.1.0 0.0.0.255	Create an ACL rule to permit certain traffic.
rule 10 deny ip	rule 10 deny ip	Create an ACL rule to deny all IP addresses except for the above permitted IP addresses.
Global settings:
acl number 3001	acl number 3001	Create an ACL.
rule 5 permit ip source 145.2.2.0 0.0.0.255 destination 146.2.2.0 0.0.0.255	rule 5 permit ip source 146.2.2.0 0.0.0.255 destination 145.2.2.0 0.0.0.255	Create an ACL rule to permit certain traffic.
rule 10 deny ip	rule 10 deny ip	Create an ACL rule to deny all IP addresses except for the above permitted IP addresses.
Interface settings:
interface FourHundredGigE1/0/1	interface FourHundredGigE2/0/1	Enter interface view.
ip address 145.1.1.1 255.255.255.0	ip address 146.1.1.1 255.255.255.0	Assign an IP address.
packet-filter 3000 inbound	packet-filter 3000 inbound	Add an inbound ACL.
interface FourHundredGigE1/0/2	interface FourHundredGigE2/0/2	Enter interface view.
ip address 145.2.2.1 255.255.255.0	ip address 146.2.2.1 255.255.255.0	Assign an IP address.
packet-filter 3001 inbound	packet-filter 3001 inbound	Add an inbound ACL.

 

Configuring VPN-based isolation

Networking configuration

As shown in Figure 7, you can configure VPN instances to meet the following isolation requirements:

·     Server1, Server2, and Server3 are isolated from each other and cannot communicate with each other.

·     Server4, Server5, and Server6 are isolated from each other and cannot communicate with each other.

·     Server1 and Server4 run in the same VPN instance and can access each other.

·     Server2 and Server5 run in the same VPN instance and can access each other.

·     Server1 and Server5 run in different VPN instances and cannot access each other.

·     Server2 and Server4 run in different VPN instances and cannot access each other.

·     Server3 and Server6 do not run in any VPN instance and can access each other.

Figure 7 Network diagram

 

Procedures

NCP1	NCP2	Description
Global settings:
ip vpn-instance vpn1	ip vpn-instance vpn1	Create VPN instance vpn1.
route-distinguisher 1:1	route-distinguisher 1:1	Specify a route distinguisher.
vpn-target 1:1 import-extcommunity	vpn-target 1:1 import-extcommunity	Specify an import route target.
vpn-target 1:1 export-extcommunity	vpn-target 1:1 export-extcommunity	Specify an export route target.
address-family ipv4	address-family ipv4	Create the IPv4 address family and enter its view.
evpn osf routing-enable	evpn osf routing-enable	Enable cell encapsulation for Type-5 routes.
address-family evpn	address-family evpn	Create the EVPN address family and enter its view.
evpn osf routing-enable	evpn osf routing-enable	Enable cell encapsulation for Type-2 routes.
ip vpn-instance vpn2	ip vpn-instance vpn2	Create VPN instance vpn2.
route-distinguisher 2:2	route-distinguisher 2:2	Specify a route distinguisher.
vpn-target 2:2 import-extcommunity	vpn-target 2:2 import-extcommunity	Specify an import route target.
vpn-target 2:2 export-extcommunity	vpn-target 2:2 export-extcommunity	Specify an export route target.
address-family ipv4	address-family ipv4	Create the IPv4 address family and enter its view.
evpn osf routing-enable	evpn osf routing-enable	Enable cell encapsulation for Type-5 routes.
address-family evpn	address-family evpn	Create the EVPN address family and enter its view.
evpn osf routing-enable	evpn osf routing-enable	Enable cell encapsulation for Type-2 routes.
ip public-instance	ip public-instance	Create the public instance and enter its view.
route-distinguisher 3:3	route-distinguisher 65535:10	Specify a route distinguisher.
vpn-target 2:1 import-extcommunity	vpn-target 2:1 import-extcommunity	Specify an import route target.
vpn-target 2:1 export-extcommunity	vpn-target 2:1 export-extcommunity	Specify an export route target.
address-family ipv4	address-family ipv4	Create the IPv4 address family and enter its view.
evpn osf routing-enable	evpn osf routing-enable	Enable cell encapsulation for Type-5 routes.
address-family evpn	address-family evpn	Create the EVPN address family and enter its view.
evpn osf routing-enable	evpn osf routing-enable	Enable cell encapsulation for Type-2 routes.
BGP instance settings:
bgp 100	bgp 100	Enter BGP instance view.
router-id 100.0.0.2	router-id 100.0.0.3	Specify a route distinguisher.
peer 100.0.0.3 as-number 100	peer 100.0.0.2 as-number 100	Create a BGP peer and specify its AS number.
address-family l2vpn evpn	address-family l2vpn evpn	Create the BGP EVPN address family and enter its view.
peer 100.0.0.3 enable	peer 100.0.0.2 enable	Enable the local router to exchange routes with the specified peer.
peer 100.0.0.3 advertise encap-type osf	peer 100.0.0.2 advertise encap-type osf	Enable the local router to advertise routes with OSF encapsulation.
Settings for BGP-VPN instance vpn1:
ip vpn-instance vpn1	ip vpn-instance vpn1	Enter BGP-VPN instance view.
peer 145.1.1.2 as-number 601	peer 146.1.1.2 as-number 701	Create a BGP peer and specify its AS number.
address-family ipv4 unicast	address-family ipv4 unicast	Create the BGP-VPN IPv4 unicast address family and enter its view.
import-route direct	import-route direct	Redistribute direct routes to BGP.
peer 145.1.1.2 enable	peer 146.1.1.2 enable	Enable the local router to exchange routes with the specified peer.
Settings for BGP-VPN instance vpn2:
ip vpn-instance vpn2	ip vpn-instance vpn2	Enter BGP-VPN instance view.
peer 145.2.2.2 as-number 602	peer 146.2.2.2 as-number 702	Create a BGP peer and specify its AS number.
address-family ipv4 unicast	address-family ipv4 unicast	Create the BGP-VPN IPv4 unicast address family and enter its view.
import-route direct	import-route direct	Redistribute direct routes to BGP.
peer 145.2.2.2 enable	peer 146.2.2.2 enable	Enable the local router to exchange routes with the specified peer.
BGP instance settings (public network):
peer 145.3.3.2 as-number 603	peer 146.3.3.2 as-number 703	Create a BGP peer and specify its AS number.
address-family ipv4 unicast	address-family ipv4 unicast	Create the BGP IPv4 unicast address family and enter its view.
import-route direct	import-route direct	Redistribute direct routes.
peer 145.3.3.2 enable	peer 146.3.3.2 enable	Enable the local router to exchange routes with the specified peer.
Interface settings:
interface FourHundredGigE 1/0/1	interface FourHundredGigE 2/0/1	Enter interface view.
ip binding vpn-instance vpn1	ip binding vpn-instance vpn1	Bind a VPN instance.
ip address 145.1.1.1 255.255.255.0	ip address 146.1.1.1 255.255.255.0	Assign an IP address.
interface FourHundredGigE 1/0/2	interface FourHundredGigE 2/0/2	Enter interface view.
ip binding vpn-instance vpn2	ip binding vpn-instance vpn2	Bind a VPN instance.
ip address 145.2.2.1 255.255.255.0	ip address 146.2.2.1 255.255.255.0	Assign an IP address.
interface FourHundredGigE 1/0/3	interface FourHundredGigE 2/0/3	Enter interface view.
ip address 145.3.3.1 255.255.255.0	ip address 146.3.3.1 255.255.255.0	Assign an IP address.