Bandwidth vs Throughput vs Latency
What you'll learn: How to distinguish between bandwidth capacity, actual throughput, and latency—three commonly confused network performance metrics.
The Three Amigos of Network Performance
When measuring network performance, these three metrics tell different but interconnected stories:
Bandwidth is the maximum capacity of your network pipe, measured in bits per second (Gbps, Mbps). Think of it as the width of a highway—how many lanes are available. A 10 Gbps Direct Connect circuit has 10 gigabits of potential capacity.
Throughput is the actual data transferred over time in real-world conditions. It's what you actually achieve, not what's theoretically possible. Using our highway analogy, throughput is how many cars actually drive through, accounting for traffic, road conditions, and accidents. Throughput is always ≤ bandwidth due to protocol overhead, packet loss, congestion, and other factors.
Latency is the time delay for data to travel from source to destination, measured in milliseconds. This is round-trip time (RTT). It's like how long a single car takes to reach its destination—regardless of highway width. A fiber connection might have 5ms latency, while a satellite connection has 600ms, even with identical bandwidth.
Why This Matters in Cloud
When troubleshooting network performance issues, you need to identify which metric is the bottleneck:
- Low throughput despite high bandwidth? Look for packet loss, TCP window issues, or application inefficiencies.
- High latency? Your data path might be geographically inefficient or traversing too many network hops.
- Insufficient bandwidth? You've hit your capacity ceiling and need to upgrade the connection size.
Key Takeaway: Bandwidth is your pipe's size, throughput is what actually flows through it, and latency is how long each bit takes to travel—optimize based on which metric limits your application performance.