SD-WAN Capacity Calculator
Estimate effective aggregate SD-WAN throughput across multiple underlying links, factoring in path quality and control-plane overhead.
Inputs
e.g. MPLS, broadband, and LTE link speeds in Mbps: 100, 50, 50
Effective usable fraction of raw link bandwidth after loss, jitter, and quality-based steering losses
Bandwidth consumed by SD-WAN control traffic, path probing, and encapsulation across all links
Effective SD-WAN Capacity
175.00Mbps
Raw Aggregate Bandwidth (sum of links)
200.00Mbps
After Path Quality Factor
180.00Mbps
Number of Links Combined
3
Control Overhead (% of raw bandwidth)
2.50%
Step by step
Values used
Link Bandwidths (comma-separated) = 100, 50, 50; Average Path Quality Factor = 90; Control-Plane Overhead = 5 Mbps
Effective SD-WAN capacity
effective = (sum of link bandwidths × path_quality_factor) − control_overhead
Effective SD-WAN Capacity
= 175.00 Mbps
Raw Aggregate Bandwidth (sum of links)
= 200.00 Mbps
After Path Quality Factor
= 180.00 Mbps
Number of Links Combined
= 3
Control Overhead (% of raw bandwidth)
= 2.50
How it works
SD-WAN combines multiple underlying transport links (MPLS, broadband, LTE/5G, satellite) into one logical connection, but the effective usable capacity is less than the simple sum of link speeds: a path quality factor accounts for real-world loss, jitter, and the SD-WAN controller steering some traffic away from degraded paths, and a fixed control-plane overhead — consumed by encapsulation, path monitoring probes, and orchestration traffic — is subtracted from the total.
Formula
Effective SD-WAN capacity
effective = (sum of link bandwidths × path_quality_factor) − control_overhead
- B_i
- Bandwidth of link i
- q
- Path quality factor (0-1)
- O
- Control-plane overhead
Frequently Asked Questions
Why isn't effective SD-WAN throughput just the sum of all link speeds?
Real-world links rarely deliver their full rated speed continuously — congestion, loss, and jitter on broadband or cellular links, plus the SD-WAN controller's own quality-based steering (which shifts traffic away from a degraded path rather than using its full nominal capacity), mean actual usable throughput runs below the simple arithmetic sum.
What consumes SD-WAN control-plane overhead?
Continuous path quality probing (measuring latency/jitter/loss on every link to make steering decisions), tunnel encapsulation headers, and orchestration/telemetry traffic to the SD-WAN controller all consume a small but constant slice of aggregate bandwidth, independent of how much user traffic is flowing.
Does adding more links always increase effective capacity proportionally?
Only if the added links have comparable quality — adding a high-latency, lossy link contributes less usable capacity per Mbps of rated speed than adding another reliable link, since the path quality factor effectively discounts unreliable paths more heavily in real deployments than this simplified average-factor model shows.