Hardware and feature specifications
Appendix A: System configuration
Appendix B: Hybrid transports VPN 0 configuration
Appendix C: Data center LAN‒side configuration
Appendix D: Quality-of-Service (QoS) configuration
Appendix E: DHCP and VRRP configuration
Appendix F: Centralized policies
Control policy applied toward branches in Group1
Application-aware routing policy for the branches
The Cisco® Software-Defined WAN (SD-WAN) is a cloud-hosted and cloud-delivered overlay WAN solution, Cisco SDWAN offers the option for those customers who desire it to host their controllers on-premise. Though hosted on premises, Cisco SD-WAN would not limit the functionalities provided through cloud-managed services, such as the time needed to deploy services, build application resiliency, and provide a robust security architecture for hybrid networks.
Cisco SD-WAN solves many critical enterprise problems, including:
● Establishing a transport-independent WAN for lower cost and higher diversity
● Meeting Service-Level Agreements (SLAs) for business-critical and real-time applications
● Providing end-to-end segmentation for protecting critical enterprise compute resources
● Extending seamlessly into the private and public cloud
● Providing secured control and data plane connectivity
Cisco SD-WAN provides data plane and control plane separation by having controllers in the private network (hosted in the data center). The diagram below shows the high-level architecture of the solution.
(The firewall ports listed below need to be opened in order for edge devices to communicate with controllers hosted within the data center network.)
Core number |
Ports for DTLS (UDP) |
Ports for TLS (TCP) |
Core0 |
12346 |
23456 |
Core1 |
12446 |
23556 |
Core2 |
12546 |
23656 |
Core3 |
12646 |
23756 |
Core4 |
12746 |
23856 |
Core5 |
12846 |
23956 |
Core6 |
12946 |
24056 |
Core7 |
13046 |
24156 |
This document covers the enterprise solution built with the features described below.
Security
The Cisco SD-WAN solution offers secure control-and-management communications between the routers and the control components. Data plane communication between the WAN edge routers is encrypted and secured based on IPsec encapsulation.
Hybrid transport
There are two data centers in this profile (DC1 and DC2), with each data center having two SD-WAN routers. All of the data centers’ SD-WAN routers are connected to the Internet and to Multiprotocol Label-Switching, Customer-Premises Equipment (MPLS CPE) routers.
The branches have a range-of-connectivity model. Some are hybrid and connected to the Internet and MPLS. Some are connected to only one transport: either to the Internet or to MPLS.
The same profile was configured and tested with dual Internet transports. Lab environment consists of Ethernet interfaces with DHCP IP address provided by service provider.
Segmentation
In the branches, there can be multiple segments; with Cisco SD-WAN, the user can keep the segments separate. In this profile, two VPN segments have been defined. One segment is used for Corporate Network (vpn10) and PCI (vpn40) for credit-card transactions, which requires flow through a firewall.
Policy-based hub-and-spoke topology
A Centralized policy is deployed to establish a hub-and-spoke topology between the data centers and the branches.
One set of branches prefers the default route from DC1, and another set of branches prefers the default from DC2.
Quality of Service (QoS)
Quality of service is configured on all devices. The WAN bandwidth is appropriately distributed between different types of applications. Voice is given dedicated bandwidth on WAN interfaces and placed in the Low Latency Queue. Other traffic classes share the remaining bandwidth among them, based on weight assignment.
App-route policies
A Centralized app-route policy is configured for hybrid sites. Voice SLAs are defined, and the MPLS is defined as the preferred path for voice traffic.
Dynamic Host Configuration Protocol (DHCP) servers for the branches
WAN edge routers in the branches are configured as DHCP servers for some of the segments, for allocating IP addresses to the clients.
High availability
In the data center, the Open Shortest Path First (OSPF) is deployed for dynamic routing.
One set of branches utilize the Virtual Router Redundancy Protocol (VRRP) on the SD-WAN edge routers connected to L2 switches within the branch. Another set of branches run Open Shortest Path First (OSPF) between the SD-WAN edge router and L3 switches within the branch.
Table 1. Profile feature summary
Deployment area |
Features |
Security |
TLS/DTLS-certificate-based control plane, IPsec-based data plane, segmentation, Zone-Based Firewall |
Services |
QoS, DIA, NAT, ACL, DHCP server |
Routing |
BGP, OSPF, VRRP |
App-aware policies |
SLA-based path selection, policy-based hub-and-spoke topology |
Centralized management |
Configuration, monitoring, and policy management through vManage |
Based on research and customer feedback and configuration samples, the Cisco SD-WAN profile is designed with a generic deployment topology that can be easily modified to fit specific deployment scenarios. This profile caters to enterprise network deployments with a large number of remote or branch offices and few data centers.
Detailed topology for remote Type A and Type B sites.
Hardware and feature specifications
This section describes the 3-D feature matrix, where the hardware platforms are listed along with their Place-In-Network (PIN), and the relevant vertical deployment.
Table 2 defines the hardware, PIN, and SD-WAN feature deployed.
Table 2. 3-D feature summary with hardware
Deployment layer (PIN) |
Platforms |
Critical vertical features |
Management plane |
3x vManage |
To scale the solution, clustering is utilized within the data-center (DC) site among the instances of vManage operating together as a single system interface from the active DC site. Within the DC site, synchronous replication among the vManage instances in the cluster is utilized to maintain states among the instances. |
Control plane |
4x vBond, 8x vSmart |
Consisting of vBond and vSmart entities, this is used to control traffic flow and policies. |
Data plane |
vEdge 1K, vEdge 2K |
Terminates the tunnels from the branches and receives and sends data packets between the branches and the data center. |
Customer Edge (CE) (MPLS circuit termination at customer edge) |
2x Cisco ASR 1006 |
Cisco ASR 1006 routers terminate AT&T and VZN 2-Gig MPLS. AT&T MPLS terminates on a Cisco ASR 1006 in both DC1 and DC2. VZN MPLS terminates on a Cisco ASR 1006 in both DC1 and DC2. Internet connectivity terminates on a Cisco Nexus® device in both GF0 and GF1. |
DC distribution layer |
4x Catalyst® 4500-x |
Catalyst 4500-x layered in pairs that work as an aggregation point for traffic between data center and headends. |
Table 3 defines the set of relevant servers, test equipment, and endpoints that are used to complete the end-to-end deployment.
This list of hardware, along with the relevant software versions and the role of these devices, complements the actual physical topology shown in Figure 1.
Table 3. Hardware profile of servers and endpoints
VM and HW |
Software version |
Description |
Spirent |
Spirent TestCenter |
Generates L7 traffic |
Ubuntu |
16.04 |
End host |
To validate a new release, the network topology is upgraded, including the new software image, with the existing configuration composed of the use cases and the relevant traffic profile. New use cases acquired from the field or from customer deployments are added on top of the existing configuration.
With respect to the longevity for this profile, the setup, the CPU, and memory use/leaks are monitored during the validation phase. Furthermore, to test the robustness of the software release and platform, negative events are triggered while executing the use cases.
Table 4 describes the use cases that were executed as a part of this profile test. These use cases are divided into buckets of technology areas to see the complete coverage of the deployment scenarios.
These technology buckets comprise System Upgrade, Security, Network Service, Monitoring & Troubleshooting, simplified management, system health monitoring, and system and network resiliency.
Table 4. List of use case scenarios
No. |
Focus area |
Use cases |
System upgrade |
||
1 |
Upgrades/downgrades |
The network administrator should be able to perform controller and vEdge upgrades (and downgrades) seamlessly between releases. All of the applied configurations should migrate seamlessly during upgrades and downgrades. |
Security |
||
2 |
ACL/IPsec/NAT |
Only authenticated devices are allowed to send traffic to one another. Provides secure communication between pairs of devices. Prevents unwanted data traffic from passing through the Viptela® vEdge routers and to the LAN networks in the service-side networks connected to the routers. |
Network services |
||
3 |
Control policy |
The network administrator is able to define the routing policies such that: For AT&T-Single MPLS locations: The routers must install a default route from GF0 data headends with higher preference 100 over a default route received from GF1 data headends with preference 50. For VZN-Single MPLS locations: The routers must install a default route from GF0 data headends with higher preference 100 over a default route received from GF1 data headends with preference 50. AT&T hubs will accept routes only from sites that are AT&T MPLS as Primary, whether single or dual. VZN headends follow the same criteria, but for VZN hubs and VZN MPLS sites. The hub should not accept a default route from any remote router at any time. |
4 |
Traffic steering policy |
The network administrator is able to steer critical vs. noncritical traffic based on the circuit to which the traffic is connected. Critical (DSCP 46, 34, 28, 26, 24, and 18) and noncritical (all other traffic): critical traffic is sent over the more reliable WAN connection while other traffic is sent over other WAN connections, for a resulting active/active path on the remote routers. For single MPLS locations: All critical traffic must traverse the MPLS while noncritical traffic takes broadband, with a possibility of failover between both connections if the SLA is violated (latency and loss). For dual MPLS where AT&T is primary: All critical traffic must traverse the AT&T MPLS while noncritical traffic takes the VZN MPLS, with a possibility of failover between both connections if the SLA is violated (latency and loss). For dual MPLS where VZN is primary: All critical traffic must traverse the VZN MPLS while noncritical traffic takes the AT&T MPLS, with a possibility of failover between both connections if the SLA is violated (latency and loss). |
5 |
Quality of Service (QoS) |
The network administrator needs to enhance the user experience by ensuring traffic and application delivery using QoS policies by classifying data packets into appropriate forwarding classes and rewriting the differentiated Services Code Point (DSCP) values. |
6 |
Application visibility |
The network administrator is able to define the application-visibility parameters so that the IPFIX information can be viewed from the collector. |
Monitoring and troubleshooting |
||
7 |
Wireshark |
The network administrator is able to troubleshoot the network by capturing and analyzing traffic. |
Simplified management |
||
8 |
Manageability |
Simplified network troubleshooting and debugging for IT administration:
●
Monitors network for alarms, syslog issues, and traps.
|
System health monitoring |
||
9 |
System health |
Monitors system health for CPU use, memory consumption, and memory leaks during testing. |
System and network resiliency and robustness |
||
10 |
System resiliency |
Verifies system-level resiliency during the following events:
●
Power failure
●
WAN/LAN interface flapping
●
Network impairments, as per SLA requirements.
|
11 |
Network resiliency |
Verifies that the system holds up well at the level of network-level resiliency. |
12 |
Negative events and triggers |
Verifies that the system holds up well and recovers to working condition after the following negative events are triggered:
●
Configuration changes: addition or removal of configuration snippets, configuration replacements routes
●
Clearing of counters, clearing of routes
●
Routing protocol interface flapping
|
Appendix A: System configuration
The system configuration is the same across all controllers and WAN edge routers.
system
host-name Spoke3
system-ip 1.1.130.1
site-id 130
admin-tech-on-failure
no route-consistency-check
sp-organization-name " esc-sdwan-dmz"
organization-name " esc-sdwan-dmz"
vbond vbonddmz.com
Appendix B: Hybrid transports VPN 0 configuration
vpn 0
dns 8.8.8.8 primary
host vbonddmz.com ip 35.164.223.65
interface ge0/0
ip address 10.151.110.1/16
nat
tunnel-interface
encapsulation ipsec
color gold restrict
allow-service all
no allow-service bgp
allow-service dhcp
allow-service dns
allow-service icmp
no allow-service sshd
no allow-service netconf
no allow-service ntp
no allow-service ospf
no allow-service stun
allow-service https
!
no shutdown
!
ip route 0.0.0.0/0 10.151.1.1
!
interface ge0/1
ip address 10.161.110.2/16
tunnel-interface
encapsulation ipsec
color mpls restrict
no allow-service bgp
allow-service dhcp
allow-service dns
allow-service icmp
no allow-service sshd
no allow-service netconf
no allow-service ntp
no allow-service ospf
no allow-service stun
allow-service https
!
no shutdown
shaping-rate 10000
qos-map WANQoS
!
ip route 0.0.0.0/0 10.161.1.1
!
ip host vbonddmz.com ip 35.164.223.65
ip name-server 8.8.4.4 8.8.8.8
ip route 0.0.0.0 0.0.0.0 10.151.1.1 1
ip route 0.0.0.0 0.0.0.0 10.161.1.1 1
interface GigabitEthernet0/0/0
no shutdown
arp timeout 1200
mtu 1500
negotiation auto
service-policy output shape_GigabitEthernet0/0/1
ip mtu 1500
ip address 10.151.130.1 255.255.0.0
exit
interface GigabitEthernet0/0/1
no shutdown
arp timeout 1200
mtu 1500
negotiation auto
service-policy output shape_GigabitEthernet0/0/0
ip mtu 1500
ip nat outside
ip address 10.161.130.1 255.255.0.0
exit
interface Tunnel0
no shutdown
ip unnumbered GigabitEthernet0/0/0
no ip redirects
ipv6 unnumbered GigabitEthernet0/0/0
no ipv6 redirects
tunnel source GigabitEthernet0/0/0
tunnel mode sdwan
exit
interface Tunnel1
no shutdown
ip unnumbered GigabitEthernet0/0/1
no ip redirects
ipv6 unnumbered GigabitEthernet0/0/1
no ipv6 redirects
tunnel source GigabitEthernet0/0/1
tunnel mode sdwan
exit
sdwan
interface GigabitEthernet0/0/0
tunnel-interface
encapsulation ipsec weight 1
color mpls restrict
no last-resort-circuit
vmanage-connection-preference 5
allow-service all
no allow-service bgp
allow-service dhcp
allow-service dns
allow-service icmp
no allow-service sshd
no allow-service netconf
no allow-service ntp
no allow-service ospf
no allow-service stun
allow-service https
exit
exit
interface GigabitEthernet0/0/1
tunnel-interface
encapsulation ipsec weight 1
color gold restrict
no last-resort-circuit
vmanage-connection-preference 5
allow-service all
no allow-service bgp
allow-service dhcp
allow-service dns
allow-service icmp
no allow-service sshd
no allow-service netconf
no allow-service ntp
no allow-service ospf
no allow-service stun
allow-service https
exit
exit
Appendix C: Data center LAN‒side configuration
vpn 10
router
ospf
router-id 1.1.20.1
timers spf 200 1000 10000
redistribute omp
area 0
interface ge0/2
exit
exit
!
!
interface ge0/2
ip address 172.16.20.2/24
no shutdown
!
!
vpn 40
router
ospf
router-id 1.1.20.2
timers spf 200 1000 10000
redistribute omp
area 0
interface ge0/3
exit
exit
!
!
interface ge0/3
ip address 172.16.24.2/24
no shutdown
!
!
vrf definition 10
rd 1:10
address-family ipv4
exit-address-family
!
address-family ipv6
exit-address-family
!
!
vrf definition 10
rd 1:40
address-family ipv4
exit-address-family
!
address-family ipv6
exit-address-family
!
!
interface GigabitEthernet1/0/0
no shutdown
arp timeout 1200
vrf forwarding 10
ip address 172.16.10.1 255.255.255.0
ip mtu 1500
ip ospf 1 area 0
ip ospf network broadcast
mtu 1500
negotiation auto
!
interface GigabitEthernet1/0/1
no shutdown
arp timeout 1200
vrf forwarding 40
ip address 172.16.14.1 255.255.255.0
ip mtu 1500
ip ospf 2 area 0
ip ospf network broadcast
mtu 1500
negotiation auto
!
router ospf 1 vrf 10
auto-cost reference-bandwidth 100
max-metric router-lsa
timers throttle spf 200 1000 10000
router-id 1.1.10.1
default-information originate
distance ospf external 110
distance ospf inter-area 110
distance ospf intra-area 110
redistribute omp subnets
!
router ospf 2 vrf 40
auto-cost reference-bandwidth 100
max-metric router-lsa
timers throttle spf 200 1000 10000
router-id 1.1.10.2
default-information originate
distance ospf external 110
distance ospf inter-area 110
distance ospf intra-area 110
redistribute omp subnets
!
Appendix D: Quality-of-Service (QoS) configuration
vpn 0
interface ge0/0
shaping-rate 10000
qos-map WANQoS
!
interface ge0/1
shaping-rate 10000
qos-map WANQoS
!
vpn 10
interface ge0/7.10
access-list LAN-Classification in
policy
class-map
class Queue0 queue 0
class Voice_EF queue 0
class Queue1 queue 1
class Queue2 queue 2
class NetProtocol_CS3 queue 3
class Queue3 queue 3
class NetMgmt_CS2 queue 4
class Queue4 queue 4
class CriticalData_AF21 queue 5
class Queue5 queue 5
class Queue6 queue 6
class Scavanger_AF11 queue 6
class BestEffort_CS1 queue 7
class Queue7 queue 7
!
access-list LAN-Classification
sequence 1
match
destination-port 1719-1721
!
action accept
class Voice_EF
set
dscp 46
!
!
!
sequence 11
match
destination-port 2326-2485
!
action accept
class Voice_EF
set
dscp 46
!
!
!
sequence 21
match
protocol 8 88 89
!
action accept
class NetProtocol_CS3
set
dscp 24
!
!
!
sequence 31
match
destination-port 22
!
action accept
class NetProtocol_CS3
set
dscp 24
!
!
!
sequence 41
match
destination-ip 10.200.200.0/24
!
action accept
class NetMgmt_CS2
set
dscp 16
!
!
!
sequence 51
match
destination-ip 10.200.201.0/24
destination-port 161 162 514
!
action accept
class CriticalData_AF21
set
dscp 20
!
!
!
sequence 61
match
destination-port 20 21
!
action accept
class BestEffort_CS1
set
dscp 8
!
!
!
sequence 71
match
destination-ip 10.200.202.0/24
!
action accept
class Scavanger_AF11
set
dscp 10
!
!
!
sequence 81
action accept
class BestEffort_CS1
set
dscp 10
!
!
!
default-action accept
!
qos-scheduler WANQoS_0
class Queue0
bandwidth-percent 11
buffer-percent 11
scheduling llq
!
qos-scheduler WANQoS_1
class Queue1
bandwidth-percent 10
buffer-percent 10
drops red-drop
!
qos-scheduler WANQoS_2
class Queue2
bandwidth-percent 10
buffer-percent 10
drops red-drop
!
qos-scheduler WANQoS_3
class Queue3
bandwidth-percent 5
buffer-percent 5
drops red-drop
!
qos-scheduler WANQoS_4
class Queue4
bandwidth-percent 2
buffer-percent 2
drops red-drop
!
qos-scheduler WANQoS_5
class Queue5
bandwidth-percent 48
buffer-percent 48
drops red-drop
!
qos-scheduler WANQoS_6
class Queue6
bandwidth-percent 5
buffer-percent 5
drops red-drop
!
qos-scheduler WANQoS_7
class Queue7
bandwidth-percent 9
buffer-percent 9
drops red-drop
!
qos-map WANQoS
qos-scheduler WANQoS_0
qos-scheduler WANQoS_1
qos-scheduler WANQoS_2
qos-scheduler WANQoS_3
qos-scheduler WANQoS_4
qos-scheduler WANQoS_5
qos-scheduler WANQoS_6
qos-scheduler WANQoS_7
!
!
Appendix E: DHCP and VRRP configuration
interface GigabitEthernet0/0/2
vrf forwarding 1
ip address 172.16.120.1 255.255.255.0
ip helper-address 172.16.10.5
negotiation auto
vrrp 1 address-family ipv4
priority 110
vrrpv2
address 172.16.120.3 primary
exit-vrrp
arp timeout 1200
vpn 1
interface ge0/2
ip address 172.16.110.2/24
dhcp-helper 172.16.10.4
no shutdown
vrrp 1
priority 110
ipv4 172.16.110.3
!
!
!
Appendix F: Centralized policies
Control policy applied toward branches in Group1
policy
control-policy PreferDC2
sequence 1
match route
site-list DC2
!
action accept
set
preference 100
!
!
!
sequence 11
match route
site-list AllBranches
vpn-list pciVPN
!
action accept
set
tloc-list DC-TLOCS
!
!
!
default-action accept
!
control-policy PreferDC1
sequence 1
match route
site-list DC1
!
action accept
set
preference 100
!
!
!
sequence 11
match route
site-list AllBranches
vpn-list pciVPN
!
action accept
set
tloc-list DC-TLOCS
!
!
!
default-action accept
!
vpn-membership vpnMembership_-258379630
sequence 10
match
vpn-list corpVPN
!
action accept
!
!
sequence 20
match
vpn-list pciVPN
!
action accept
!
!
default-action reject
!
data-policy _guestVPN_Drop1918
vpn-list guestVPN
sequence 1
match
destination-data-prefix-list RFC1918Plus
!
action accept
!
!
default-action accept
!
lists
data-prefix-list RFC1918Plus
ip-prefix 10.0.0.0/8
ip-prefix 172.16.0.0/16
ip-prefix 192.168.0.0/16
ip-prefix 198.18.128.0/18
!
site-list AllBranches
site-id 300-499
!
site-list BranchG1
site-id 300-399
!
site-list BranchG2
site-id 400-499
!
site-list DC1
site-id 100
!
site-list DC2
site-id 200
!
tloc-list DC-TLOCS
tloc 10.1.0.1 color mpls encap ipsec
tloc 10.1.0.1 color biz-internet encap ipsec
tloc 10.1.0.2 color mpls encap ipsec
tloc 10.1.0.2 color biz-internet encap ipsec
tloc 10.2.0.1 color mpls encap ipsec
tloc 10.2.0.1 color biz-internet encap ipsec
tloc 10.2.0.2 color mpls encap ipsec
tloc 10.2.0.2 color biz-internet encap ipsec
!
vpn-list corpVPN
vpn 10
!
vpn-list guestVPN
vpn 40
!
vpn-list pciVPN
vpn 20
!
!
!
apply-policy
site-list BranchG1
control-policy PreferDC1 out
!
site-list AllBranches
data-policy _guestVPN_Drop1918 from-service
vpn-membership vpnMembership_-258379630
!
site-list BranchG2
control-policy PreferDC2 out
!
!
Application-aware routing policy for the branches
policy
sla-class BestEffort
latency 250
loss 10
jitter 30
!
sla-class CriticalData
latency 200
loss 3
jitter 20
!
sla-class Voice
latency 150
loss 1
jitter 5
!
app-route-policy _storeVPN_CVP-APP-Route1
vpn-list storeVPN
sequence 1
match
dscp 46
!
action
sla-class Voice preferred-color mpls
!
!
sequence 11
match
dscp 20
!
action
sla-class CriticalData preferred-color mpls
!
!
sequence 21
match
dscp 0-10
!
action
sla-class BestEffort preferred-color gold
!
!
!
lists
prefix-list DefaultPrefix
ip-prefix 0.0.0.0/0
!
site-list BranchGroup1
site-id 1000-1999
!
site-list BranchGroup2
site-id 2000-2999
!
site-list DC1
site-id 100
!
site-list DC2
site-id 200
!
vpn-list storeVPN
vpn 10
!
!
!
apply-policy
site-list BranchGroup1
control-policy Group1BranchControl-Out out
!
!