2.10 High Availability Design
| Field |
Value |
| Document Title |
High Availability Design |
| Section Number |
2.10 |
| Version |
1.0 |
| Last Updated |
December 30, 2025 |
| Author |
Network Architecture Team |
| Classification |
Internal Use |
2.10.1 HA Architecture Overview
Availability Requirements
| Component |
Target Availability |
Max Downtime/Year |
Recovery Time Objective |
| Control Plane |
99.999% |
5.26 minutes |
<30 seconds |
| Data Plane |
99.99% |
52.6 minutes |
<5 seconds |
| Management Plane |
99.9% |
8.76 hours |
<15 minutes |
| Overall SD-WAN |
99.99% |
52.6 minutes |
<60 seconds |
HA Design Principles
HA DESIGN PRINCIPLES
+==============================================================================+
| |
| 1. NO SINGLE POINT OF FAILURE |
| +------------------------------------------------------------------+ |
| | - Dual controllers per region | |
| | - Dual WAN edges at hub sites | |
| | - Multiple transport paths | |
| | - Geo-redundant management | |
| +------------------------------------------------------------------+ |
| |
| 2. GRACEFUL DEGRADATION |
| +------------------------------------------------------------------+ |
| | - Continue operations during controller failure | |
| | - Maintain forwarding during management failure | |
| | - Automatic path failover during transport failure | |
| +------------------------------------------------------------------+ |
| |
| 3. FAST CONVERGENCE |
| +------------------------------------------------------------------+ |
| | - BFD for sub-second detection | |
| | - Pre-computed backup paths | |
| | - Stateful failover where possible | |
| +------------------------------------------------------------------+ |
| |
| 4. CONSISTENT POLICY |
| +------------------------------------------------------------------+ |
| | - Policy sync across controllers | |
| | - Configuration persistence on WAN edges | |
| | - Graceful restart during upgrades | |
| +------------------------------------------------------------------+ |
| |
+==============================================================================+
2.10.2 Controller HA Design
vManage Cluster Architecture
vMANAGE CLUSTER HA ARCHITECTURE
+==============================================================================+
| |
| MUMBAI DC (PRIMARY) CHENNAI DR (DISASTER RECOVERY) |
| +---------------------------+ +---------------------------+ |
| | | | | |
| | +-------+ +-------+ | | +-------+ +-------+ | |
| | |vMng-1 | |vMng-2 | | | |vMng-4 | |vMng-5 | | |
| | |Active | |Active | | | |Standby| |Standby| | |
| | +---+---+ +---+---+ | | +---+---+ +---+---+ | |
| | | | | | | | | |
| | +----+----+ | | +----+----+ | |
| | | | | | | |
| | +----+----+ | | +----+----+ | |
| | |vMng-3 | | | |vMng-6 | | |
| | |Active | | | |Standby | | |
| | +---------+ | | +---------+ | |
| | | | | |
| | Cluster Network: | | Cluster Network: | |
| | 192.168.1.0/24 | Async | 192.168.2.0/24 | |
| | | Repl. | | |
| | Services: |<========>| Services: | |
| | - Config DB (Raft) | 15 min | - Config DB (Replica) | |
| | - Stats DB (Distrib) | RPO | - Stats DB (Replica) | |
| | - Application Server | | - Application Server | |
| | - NMS | | - NMS | |
| | | | | |
| +---------------------------+ +---------------------------+ |
| |
| FAILOVER: |
| - Intra-cluster: Automatic (<30 sec), handled by Raft consensus |
| - Inter-site (DR): Manual activation, RPO 15 min, RTO 30 min |
| |
+==============================================================================+
vSmart Controller HA
| Location |
Primary vSmart |
Secondary vSmart |
Affinity |
| Mumbai DC |
vSmart-1 (10.255.255.1) |
vSmart-2 (10.255.255.2) |
India Region |
| Chennai DR |
vSmart-3 (10.255.255.3) |
vSmart-4 (10.255.255.4) |
EMEA/Americas |
vSmart Failover Behavior:
!======================================================================
! vSMART HA BEHAVIOR
!======================================================================
!
! Normal Operation:
! - WAN Edges connect to both vSmarts in their affinity group
! - OMP sessions active to all reachable vSmarts
! - Routes learned from all vSmarts (best path selection)
!
! Failover Scenario (vSmart-1 failure):
! - BFD detects failure in <3 seconds
! - WAN Edges continue with vSmart-2
! - OMP graceful restart preserves routes for 12 hours
! - No traffic disruption (data plane independent)
!
! Recovery:
! - vSmart-1 rejoins cluster
! - OMP sessions re-established
! - Route sync automatic
!
! Configuration:
omp
graceful-restart
timers
graceful-restart-timer 43200
holdtime 60
advertisement-interval 1
vBond Orchestrator HA
vBOND HA ARCHITECTURE
+==============================================================================+
| |
| CLOUD-HOSTED vBOND (ACTIVE-ACTIVE) |
| |
| +---------------------------+ +---------------------------+ |
| | AWS ap-south-1 | | AWS eu-west-1 | |
| | (Mumbai Region) | | (Ireland Region) | |
| | | | | |
| | +-------------------+ | | +-------------------+ | |
| | | vBond-1 | | | | vBond-2 | | |
| | | 13.234.xxx.xxx | | | | 52.214.xxx.xxx | | |
| | | c5.large | | | | c5.large | | |
| | +-------------------+ | | +-------------------+ | |
| | | | | |
| +---------------------------+ +---------------------------+ |
| | | |
| +----------------+-----------------+ |
| | |
| +------+------+ |
| | Route 53 | |
| | Latency DNS | |
| | vbond. | |
| | abhavtech. | |
| | com | |
| +-------------+ |
| |
| DNS RESOLUTION: |
| - India/APAC clients -> vBond-1 (Mumbai) |
| - EMEA/Americas clients -> vBond-2 (Ireland) |
| - Automatic failover via Route 53 health checks |
| |
+==============================================================================+
2.10.3 WAN Edge HA Design
Site HA Models
| Site Type |
HA Model |
Devices |
Failover Time |
| Data Center |
Active-Active |
2x C8500 |
<1 second |
| Regional Hub |
Active-Active |
2x C8300 |
<1 second |
| Large Branch |
Active-Standby |
2x C8200L |
<5 seconds |
| Small Branch |
Single Device |
1x C8200L |
N/A (transport redundancy) |
Active-Active Configuration (Hub Sites)
ACTIVE-ACTIVE WAN EDGE HA
+==============================================================================+
| |
| MUMBAI DC - ACTIVE-ACTIVE |
| |
| +------------------------+ +------------------------+ |
| | WAN EDGE 1 | | WAN EDGE 2 | |
| | C8500-12X4QC | | C8500-12X4QC | |
| | 10.1.100.1 | | 10.1.100.2 | |
| | | | | |
| | Transport: | | Transport: | |
| | - MPLS (Te0/0/0) | | - MPLS (Te0/0/0) | |
| | - DIA (Te0/0/1) | | - DIA (Te0/0/1) | |
| | - 5G (Cell0/4/0) | | - 5G (Cell0/4/0) | |
| | | | | |
| | Service VPNs: | | Service VPNs: | |
| | - VPN 10 (Corporate) | | - VPN 10 (Corporate) | |
| | - VPN 20 (Guest) | | - VPN 20 (Guest) | |
| | - VPN 40 (Voice) | | - VPN 40 (Voice) | |
| | | | | |
| +----------+-------------+ +----------+-------------+ |
| | | |
| | ECMP LOAD BALANCING | |
| +----------------+-----------------+ |
| | |
| +------+------+ |
| | L3 Switch | |
| | (Core) | |
| +-------------+ |
| |
| TRAFFIC DISTRIBUTION: |
| - 50/50 load sharing via ECMP |
| - Per-flow load balancing |
| - Independent tunnel sets per device |
| - BFD monitors both paths |
| |
+==============================================================================+
Active-Standby Configuration (Branch Sites)
!======================================================================
! ACTIVE-STANDBY WAN EDGE CONFIGURATION - BANGALORE
!======================================================================
!
! Primary WAN Edge (Active)
system
system-ip 10.1.103.1
site-id 103
organization-name ABHAVTECH-COM
!
! TLOC Preference (Higher = Preferred)
interface GigabitEthernet0/0/0
sdwan
tunnel-interface
color biz-internet
encapsulation ipsec weight 1
tloc-extension GigabitEthernet0/0/1
!
! Secondary WAN Edge (Standby)
system
system-ip 10.1.103.2
site-id 103
organization-name ABHAVTECH-COM
!
interface GigabitEthernet0/0/0
sdwan
tunnel-interface
color biz-internet
encapsulation ipsec weight 10 ! Lower preference
!
! VRRP for LAN Gateway
interface GigabitEthernet0/0/1
ip address 10.103.1.2 255.255.255.0
vrrp 1 ip 10.103.1.1
vrrp 1 priority 110
vrrp 1 preempt delay minimum 60
vrrp 1 track 1 decrement 20
!
track 1 interface GigabitEthernet0/0/0 line-protocol
2.10.4 Transport HA Design
Multi-Transport Redundancy
TRANSPORT REDUNDANCY MATRIX
+==============================================================================+
| |
| SITE | TRANSPORT 1 | TRANSPORT 2 | TRANSPORT 3 |
| --------------+----------------+----------------+------------------------- |
| Mumbai DC | MPLS (Primary) | DIA (Active) | 5G (Last Resort) |
| Chennai DR | MPLS (Primary) | DIA (Active) | 5G (Last Resort) |
| London | MPLS (Primary) | DIA (Active) | 5G (Last Resort) |
| Frankfurt | MPLS (Primary) | DIA (Active) | 5G (Last Resort) |
| New Jersey | MPLS (Primary) | DIA (Active) | 5G (Last Resort) |
| Dallas | MPLS (Primary) | DIA (Active) | 5G (Last Resort) |
| Bangalore | DIA (Primary) | LTE (Backup) | - |
| Delhi | DIA (Primary) | LTE (Backup) | - |
| Noida | DIA (Primary) | LTE (Backup) | - |
| |
| FAILOVER BEHAVIOR: |
| +------------------------------------------------------------------+ |
| | Trigger | Detection | Action | Time | |
| +------------------------------------------------------------------+ |
| | MPLS failure | BFD (3 sec) | Shift to DIA | <5 sec | |
| | DIA failure | BFD (3 sec) | Shift to MPLS/5G | <5 sec | |
| | Dual failure | BFD (3 sec) | Activate LTE/5G | <10 sec | |
| | Quality degrade | AAR (60 sec) | Path switch | <90 sec | |
| +------------------------------------------------------------------+ |
| |
+==============================================================================+
BFD Configuration for Fast Failover
!======================================================================
! BFD CONFIGURATION FOR TRANSPORT HA
!======================================================================
!
! BFD for MPLS Transport (Tight timers - reliable transport)
interface TenGigabitEthernet0/0/0
sdwan
tunnel-interface
color mpls
encapsulation ipsec
bfd
color mpls
hello-interval 1000
multiplier 3
pmtu-discovery
!
! BFD for Internet Transport (Standard timers)
interface TenGigabitEthernet0/0/1
sdwan
tunnel-interface
color biz-internet
encapsulation ipsec
bfd
color biz-internet
hello-interval 1000
multiplier 7
pmtu-discovery
!
! BFD for Cellular Backup (Relaxed timers - variable latency)
interface Cellular0/4/0
sdwan
tunnel-interface
color 5g
encapsulation ipsec
last-resort-circuit
bfd
color 5g
hello-interval 1000
multiplier 10
pmtu-discovery
2.10.5 Application-Level HA
Application-Aware Routing HA
AAR-BASED APPLICATION HA
+==============================================================================+
| |
| APPLICATION SLA CLASSES |
| +------------------------------------------------------------------+ |
| | Class | Latency | Jitter | Loss | Failover Behavior | |
| +------------------------------------------------------------------+ |
| | Real-Time | <150ms | <30ms | <1% | Immediate switch | |
| | Business-Crit | <200ms | <50ms | <2% | Switch within 60s | |
| | Default | <300ms | <100ms | <5% | Switch within 120s | |
| | Best-Effort | N/A | N/A | N/A | No automatic switch | |
| +------------------------------------------------------------------+ |
| |
| EXAMPLE: Voice Traffic (Real-Time SLA) |
| |
| Normal: MPLS (Latency: 8ms, Loss: 0.01%) |
| | |
| v |
| Event: MPLS latency spikes to 200ms |
| | |
| v |
| AAR Action: Immediate failover to DIA (Latency: 15ms) |
| | |
| v |
| Recovery: When MPLS recovers, traffic returns (after 10 min stability) |
| |
+==============================================================================+
SLA Class Configuration
!======================================================================
! APPLICATION SLA CLASSES FOR HA
!======================================================================
!
! Real-Time SLA Class (Voice/Video)
policy
sla-class Real-Time
latency 150
jitter 30
loss 1
!
! Business-Critical SLA Class
policy
sla-class Business-Critical
latency 200
jitter 50
loss 2
!
! App-Route Policy for Voice
policy
app-route-policy VOICE-HA
vpn-list VPN-40
sequence 10
match
app-list VOICE-APPS
action
sla-class Real-Time preferred-color mpls
sla-class Real-Time strict
backup-sla-preferred-color biz-internet
!
! Fallback Behavior
policy
app-route-policy VOICE-HA
vpn-list VPN-40
sequence 20
match
app-list VOICE-APPS
action
sla-class Real-Time preferred-color biz-internet
fallback-to-best-path
2.10.6 Site-Level DR Design
Disaster Recovery Architecture
SITE-LEVEL DISASTER RECOVERY
+==============================================================================+
| |
| NORMAL OPERATION |
| +------------------------------------------------------------------+ |
| | | |
| | MUMBAI DC (Primary) CHENNAI DR (Standby) | |
| | +-----------------+ +-----------------+ | |
| | | vManage Cluster | Async | vManage Cluster | | |
| | | (Active) | Repl. | (Standby) | | |
| | +-----------------+ =====> +-----------------+ | |
| | | |
| | +-----------------+ +-----------------+ | |
| | | WAN Edges | Active | WAN Edges | | |
| | | (Primary Exit) | Tunnels | (Backup Exit) | | |
| | +-----------------+ +-----------------+ | |
| | | |
| +------------------------------------------------------------------+ |
| |
| DR ACTIVATION (Mumbai DC Failure) |
| +------------------------------------------------------------------+ |
| | | |
| | 1. Detect Mumbai failure (BFD/heartbeat timeout) | |
| | 2. Activate Chennai vManage cluster (manual or auto) | |
| | 3. Update DNS for vManage VIP | |
| | 4. WAN Edges reconnect to Chennai controllers | |
| | 5. Traffic reroutes through Chennai WAN Edges | |
| | 6. RPO: 15 minutes (config), RTO: 30 minutes | |
| | | |
| +------------------------------------------------------------------+ |
| |
+==============================================================================+
DR Runbook Summary
| Step |
Action |
Owner |
Time |
| 1 |
Confirm Mumbai DC outage |
NOC |
5 min |
| 2 |
Declare DR event |
Management |
5 min |
| 3 |
Activate Chennai vManage |
Network Team |
10 min |
| 4 |
Update DNS records |
Network Team |
5 min |
| 5 |
Verify controller connectivity |
Network Team |
5 min |
| 6 |
Validate traffic flow |
NOC |
10 min |
| Total |
|
|
40 min |
2.10.7 HA Verification and Testing
HA Verification Commands
!======================================================================
! HA VERIFICATION COMMANDS
!======================================================================
!
! vManage Cluster Status
show running-config cluster
show cluster status
!
! vSmart Controller Status
show sdwan control connections
show sdwan omp peers
!
! WAN Edge HA Status
show sdwan bfd sessions
show sdwan tunnel statistics
show vrrp
!
! Transport Health
show sdwan app-route statistics
show sdwan policy app-route-policy-filter
HA Testing Schedule
| Test Type |
Frequency |
Duration |
Impact |
| vManage node failover |
Quarterly |
2 hours |
None (cluster HA) |
| vSmart failover |
Quarterly |
1 hour |
None (graceful restart) |
| WAN Edge failover |
Monthly |
30 min |
Sub-second traffic shift |
| Transport failover |
Monthly |
15 min |
<5 second traffic shift |
| Full DR test |
Annually |
8 hours |
Planned maintenance window |
Document Control
| Version |
Date |
Author |
Changes |
| 1.0 |
December 30, 2025 |
Network Architecture Team |
Initial release |
This document is part of the Abhavtech.com SD-WAN Documentation Suite
Confidential - Internal Use Only