2.11 Sizing & Scalability Design¶

Document Information¶

Field	Value
Document Title	SD-WAN Sizing & Scalability Design
Version	1.0
Author	Network Architecture Team
Organization	Abhavtech.com
Last Updated	December 2025
Status	Production

Table of Contents¶

Sizing Overview
Controller Sizing
WAN Edge Sizing
Bandwidth Calculations
Performance Baselines
Scalability Planning
Growth Projections
Capacity Monitoring

1. Sizing Overview¶

1.1 Sizing Methodology¶

The sizing approach follows Cisco's validated design principles with safety margins for growth.

Sizing Factors: - Number of WAN edges and sites - Aggregate throughput requirements - Tunnel count and complexity - Feature enablement (encryption, UTD, DPI) - Policy complexity and rule count - Template and configuration size - Logging and analytics volume

1.2 Current Deployment Scale¶

Metric	Current	Year 1	Year 3	Year 5
Sites	9	12	25	50
WAN Edges	15	22	45	85
Users	2,500	3,500	6,000	10,000
Bandwidth (Mbps)	4,500	7,000	15,000	30,000
VPNs/Segments	5	8	12	15
Policies	50	100	200	350

2. Controller Sizing¶

2.1 SD-WAN Manager (vManage) Cluster¶

Cluster Configuration:

+-------------------------------------------------------------------+
|                    SD-WAN MANAGER CLUSTER                          |
+-------------------------------------------------------------------+
|                                                                    |
|  Mumbai DC (Primary)           Chennai DR (Standby)                |
|  +-------------------+         +-------------------+               |
|  | vManage-1 (Master)|         | vManage-DR1      |               |
|  | 32 vCPU, 128GB    |   ----> | 32 vCPU, 128GB   |               |
|  | 2TB SSD RAID10    |  Async  | 2TB SSD RAID10   |               |
|  +-------------------+  Repl   +-------------------+               |
|  | vManage-2         |         | vManage-DR2      |               |
|  | 32 vCPU, 128GB    |   ----> | 32 vCPU, 128GB   |               |
|  | 2TB SSD RAID10    |         | 2TB SSD RAID10   |               |
|  +-------------------+         +-------------------+               |
|  | vManage-3         |         | vManage-DR3      |               |
|  | 32 vCPU, 128GB    |   ----> | 32 vCPU, 128GB   |               |
|  | 2TB SSD RAID10    |         | 2TB SSD RAID10   |               |
|  +-------------------+         +-------------------+               |
|                                                                    |
|  Cluster VIP: 10.254.1.100    DR VIP: 10.254.2.100               |
+-------------------------------------------------------------------+

vManage Server Specifications:

Scale	vCPU	Memory	Storage	Devices	Notes
Small (1-100)	16	64 GB	500 GB	Up to 100	Single node
Medium (100-1000)	24	96 GB	1 TB	Up to 1,000	3-node cluster
Large (1000-5000)	32	128 GB	2 TB	Up to 5,000	6-node cluster
Abhavtech	32	128 GB	2 TB	Target: 85	3+3 cluster

Storage Requirements:

Component	Size	Purpose
OS & Software	100 GB	Base installation
Configuration DB	200 GB	Device configs, templates
Statistics DB	500 GB	Performance metrics (30 days)
Logs	300 GB	System and audit logs
Backups	500 GB	Configuration backups
Growth Buffer	400 GB	50% growth allowance
Total	2 TB	Per node

2.2 SD-WAN Controller (vSmart) Sizing¶

vSmart Specifications:

Parameter	Value	Notes
vCPU	8	Per controller
Memory	16 GB	OMP table capacity
Storage	100 GB	SSD recommended
Deployment	4 controllers	2 Mumbai, 2 Chennai
Edge Capacity	5,000 per vSmart	Supports 20,000 total
OMP Routes	128,000 per vSmart	Regional summarization

vSmart Distribution:

Mumbai DC                        Chennai DR
+----------------+               +----------------+
| vSmart-1       |               | vSmart-3       |
| Region 1+3     |               | Region 2+3     |
| 10.254.1.20    |               | 10.254.2.20    |
+----------------+               +----------------+
| vSmart-2       |               | vSmart-4       |
| Region 1+3     |               | Region 2+3     |
| 10.254.1.21    |               | 10.254.2.21    |
+----------------+               +----------------+

2.3 SD-WAN Validator (vBond) Sizing¶

vBond Specifications:

Parameter	Value
vCPU	4
Memory	8 GB
Storage	50 GB
Deployment	2 (cloud-hosted)
Edge Capacity	20,000 per vBond

3. WAN Edge Sizing¶

3.1 Platform Selection Matrix¶

Site Type	Model	Throughput	Sessions	Tunnels	Sites
Data Center	C8500-12X4QC	10 Gbps	2M	8,000	Mumbai, Chennai
Regional Hub	C8300-2N2S-6T	2.5 Gbps	500K	4,000	London, Frankfurt, NJ, Dallas
Large Branch	C8200-1N-4T	1 Gbps	250K	2,000	Bangalore, Delhi
Small Branch	C8200L-1N-4T	500 Mbps	100K	1,000	Noida

3.2 Throughput Requirements by Site¶

Mumbai DC (Primary Data Center):

Traffic Type	Current (Mbps)	Year 3 (Mbps)	Year 5 (Mbps)
Branch-to-DC	1,500	3,500	7,000
Inter-DC	800	2,000	4,000
Cloud (SaaS)	500	1,500	3,000
Cloud (IaaS)	300	1,000	2,000
Total	3,100	8,000	16,000
Platform	C8500-12X4QC (10G)	C8500-12X4QC	2x C8500-12X4QC

Regional Hub (London/Frankfurt/NJ/Dallas):

Traffic Type	Current (Mbps)	Year 3 (Mbps)	Year 5 (Mbps)
Local Users	200	400	800
Regional Branches	300	600	1,200
Cloud Access	200	500	1,000
Total	700	1,500	3,000
Platform	C8300-2N2S-6T (2.5G)	C8300-2N2S-6T	C8300-2N2S-6T

Branch Site (Bangalore/Delhi):

Traffic Type	Current (Mbps)	Year 3 (Mbps)	Year 5 (Mbps)
Business Apps	150	300	500
Cloud/SaaS	100	250	400
Voice/Video	50	100	200
Total	300	650	1,100
Platform	C8200-1N-4T (1G)	C8200-1N-4T	C8300-2N2S-6T

3.3 Feature Impact on Performance¶

Throughput Reduction Factors:

Feature	Impact	Mitigation
IPsec AES-256	10-15%	Hardware crypto engine
UTD IPS/IDS	30-40%	Size for encrypted throughput
DPI	20-30%	Limit to critical applications
Application Visibility	5-10%	Enable selectively
SSL Inspection	40-50%	SSL proxy for specific traffic

Effective Throughput Calculation:

Effective_Throughput = Raw_Throughput × (1 - IPsec_Impact) × (1 - UTD_Impact)

Example (C8300 with IPsec + UTD):
  Raw: 2,500 Mbps
  IPsec: 2,500 × 0.85 = 2,125 Mbps
  UTD: 2,125 × 0.65 = 1,381 Mbps effective

4. Bandwidth Calculations¶

4.1 Site Bandwidth Requirements¶

Bandwidth Planning Formula:

Total_BW = (Users × BW_per_User) + (Apps × App_BW) + Overhead(20%)

Mumbai DC:
  Users: 800 × 2 Mbps = 1,600 Mbps
  Apps: 15 × 100 Mbps = 1,500 Mbps
  Overhead: 620 Mbps
  Total: 3,720 Mbps → Provision 5 Gbps

4.2 Aggregate Bandwidth Summary¶

Site	Users	Calculated BW	Provisioned	Transport Mix
Mumbai DC	800	3,720 Mbps	5,000 Mbps	MPLS 2G + DIA 2G + 5G 1G
Chennai DR	600	2,790 Mbps	4,000 Mbps	MPLS 2G + DIA 1.5G + 5G 500M
Bangalore	400	1,860 Mbps	2,000 Mbps	MPLS 500M + DIA 1G + LTE 500M
Delhi	350	1,627 Mbps	2,000 Mbps	MPLS 500M + DIA 1G + LTE 500M
Noida	100	465 Mbps	500 Mbps	DIA 300M + LTE 200M
London	150	697 Mbps	1,000 Mbps	MPLS 500M + DIA 500M
Frankfurt	50	232 Mbps	500 Mbps	MPLS 200M + DIA 300M
New Jersey	30	139 Mbps	500 Mbps	MPLS 200M + DIA 300M
Dallas	20	93 Mbps	300 Mbps	DIA 200M + LTE 100M

4.3 Control Plane Bandwidth¶

OMP/BFD Overhead:

Traffic Type	Bandwidth	Notes
OMP Updates	50-100 Kbps	Per vSmart connection
BFD (per tunnel)	5 Kbps	100ms intervals
DTLS Keepalives	10 Kbps	Per controller
Statistics Upload	200-500 Kbps	Per edge to vManage
Total per Edge	~1 Mbps	Control plane reservation

5. Performance Baselines¶

5.1 Tunnel Performance Targets¶

Metric	Target	Warning	Critical
Tunnel Latency	<150 ms	>150 ms	>200 ms
Jitter	<30 ms	>30 ms	>50 ms
Packet Loss	<0.5%	>0.5%	>1%
Tunnel Flaps	0/day	>2/day	>5/day

5.2 Application Performance SLAs¶

Application Class	Latency	Jitter	Loss	Bandwidth
Real-Time (Voice)	<150 ms	<30 ms	<1%	100 Kbps/call
Interactive (Video)	<200 ms	<50 ms	<2%	2 Mbps/stream
Business Critical	<300 ms	<100 ms	<3%	Variable
Default	<500 ms	<150 ms	<5%	Best effort

5.3 Controller Performance Targets¶

Metric	vManage	vSmart	vBond
CPU Utilization	<70%	<60%	<50%
Memory Utilization	<80%	<70%	<60%
API Response Time	<2 sec	N/A	N/A
Config Push Time	<30 sec	N/A	N/A
OMP Convergence	N/A	<10 sec	N/A

6. Scalability Planning¶

6.1 Scalability Limits¶

Maximum Supported Scale:

Component	Limit	Abhavtech Current	Headroom
Sites	20,000	9	99.95%
WAN Edges	20,000	15	99.92%
Service VPNs	512	5	99.0%
Tunnels per Edge	8,000	189	97.6%
OMP Routes	512,000	~2,000	99.6%
Policies	10,000	50	99.5%
Templates	2,000	25	98.8%

6.2 Scalability Architecture¶

+------------------------------------------------------------------+
|                    SCALABILITY ARCHITECTURE                       |
+------------------------------------------------------------------+
|                                                                   |
|  Tier 1: Regional Edges (Branch/Small Sites)                     |
|  +---------------------------------------------------------+     |
|  | 85 WAN Edges (Year 5) | C8200/C8300 | OMP to vSmart     |     |
|  +---------------------------------------------------------+     |
|                              |                                    |
|  Tier 2: Hub Sites (Aggregation)                                 |
|  +---------------------------------------------------------+     |
|  | 8 Hub Edges | C8300/C8500 | Route Summarization         |     |
|  +---------------------------------------------------------+     |
|                              |                                    |
|  Tier 3: Controllers                                             |
|  +---------------------------------------------------------+     |
|  | 4 vSmart | Regional Distribution | MRF Enabled          |     |
|  +---------------------------------------------------------+     |
|                              |                                    |
|  Tier 4: Management                                              |
|  +---------------------------------------------------------+     |
|  | 6 vManage (3+3) | Clustered | Geo-Redundant             |     |
|  +---------------------------------------------------------+     |
+------------------------------------------------------------------+

6.3 Scale-Out Triggers¶

Metric	Current	Scale-Out Trigger	Action
WAN Edges	15	>100 edges	Add vSmart pair
OMP Routes	~2,000	>50,000 routes	Enable route summarization
vManage CPU	35%	>70% sustained	Add cluster node
Templates	25	>500 templates	Implement hierarchy
Statistics Volume	10 GB/day	>50 GB/day	Add stats node

7. Growth Projections¶

7.1 5-Year Growth Model¶

Sites Growth:
Year 0: ████████░░░░░░░░░░░░ 9 sites (Current)
Year 1: ██████████░░░░░░░░░░ 12 sites (+33%)
Year 2: ████████████░░░░░░░░ 18 sites (+50%)
Year 3: ██████████████░░░░░░ 25 sites (+39%)
Year 4: ████████████████░░░░ 35 sites (+40%)
Year 5: ██████████████████░░ 50 sites (+43%)

WAN Edges Growth:
Year 0: 15 edges   → Year 5: 85 edges (467% growth)
Year 0: 2,500 users → Year 5: 10,000 users (300% growth)

7.2 Capacity Planning Timeline¶

Year	Action	Investment
Year 0	Initial deployment	Controllers + 15 edges
Year 1	Regional expansion	+7 edges, enhanced licensing
Year 2	India expansion	+15 edges, vSmart addition
Year 3	APAC expansion	+10 edges, storage expansion
Year 4	Americas growth	+15 edges, vManage upgrade
Year 5	Global reach	+20 edges, platform refresh

7.3 Resource Forecasting¶

Controller Capacity Forecast:

Year	Edges	vManage CPU	vManage Memory	vSmart OMP Routes
0	15	35%	40%	2,000
1	22	40%	45%	4,000
2	37	50%	55%	8,000
3	47	55%	60%	12,000
4	65	65%	70%	18,000
5	85	75%	80%	25,000

8. Capacity Monitoring¶

8.1 Key Capacity Metrics¶

vManage Monitoring:

# Check cluster status
show cluster status

# Monitor resource utilization
show system status

# Check configuration database
show configuration database status

# Statistics database health
show statistics database status

vSmart Monitoring:

# OMP peer count
show sdwan omp peers | count

# OMP route count
show sdwan omp routes | count

# Controller resources
show system status

8.2 Capacity Dashboard Metrics¶

Category	Metric	Query Interval
Controllers	CPU/Memory/Disk	5 min
WAN Edges	Throughput/Sessions	1 min
Tunnels	Count/Utilization	5 min
OMP	Route Count/Churn	15 min
Policies	Applied/Pending	30 min
Licenses	Usage/Available	Daily

8.3 Alerting Thresholds¶

Resource	Warning	Critical	Action
CPU	70%	85%	Scale out or optimize
Memory	75%	90%	Add capacity
Disk	70%	85%	Expand storage
Tunnels	80% max	90% max	Add WAN edge
OMP Routes	80% limit	90% limit	Enable summarization
Licenses	80% used	95% used	Procure additional

8.4 Capacity Planning Automation¶

Python Script for Capacity Forecasting:

#!/usr/bin/env python3
"""
SD-WAN Capacity Forecasting Script
Abhavtech.com
"""

import requests
from datetime import datetime, timedelta

def get_current_capacity(vmanage_host, auth):
    """Retrieve current capacity metrics"""
    metrics = {}

    # Get edge count
    response = requests.get(
        f"https://{vmanage_host}/dataservice/device",
        auth=auth, verify=False
    )
    metrics['edge_count'] = len(response.json()['data'])

    # Get OMP route count
    response = requests.get(
        f"https://{vmanage_host}/dataservice/statistics/omp/routes",
        auth=auth, verify=False
    )
    metrics['omp_routes'] = response.json()['totalCount']

    return metrics

def forecast_capacity(current, growth_rate, years):
    """Project capacity needs"""
    projections = []
    for year in range(years + 1):
        projected = current * ((1 + growth_rate) ** year)
        projections.append({
            'year': year,
            'value': int(projected)
        })
    return projections

# Usage
current_edges = 15
growth_rate = 0.40  # 40% annual growth
forecast = forecast_capacity(current_edges, growth_rate, 5)

Summary¶

The sizing and scalability design ensures Abhavtech's SD-WAN infrastructure can support current requirements while accommodating projected 5-year growth from 9 to 50 sites.

Key Sizing Decisions: - vManage: 3+3 cluster (32 vCPU, 128 GB, 2 TB each) - vSmart: 4 controllers geo-distributed - WAN Edges: C8500 (DC), C8300 (hub), C8200 (branch) - Bandwidth: 20% overhead, feature impact factors applied - Growth: 40% annual expansion capacity built-in

Next Section: 2.12 IP Addressing Design

Document Version: 1.0 Last Updated: December 2025 Classification: Internal Use