About the Bandwidth-Delay Product
The Bandwidth-Delay Product (BDP) tells you how much data can be "in flight" on a network link at any instant — the amount of unacknowledged data needed to keep a connection running at full speed. It is the starting point for sizing TCP send and receive buffers, especially on long-distance or high-speed links such as satellite, transcontinental fiber, or fast WAN circuits.
Understanding the formula
BDP = Bandwidth × Round-Trip Time. Bandwidth is the link's data rate in bits per second, and RTT is the time for a packet to travel to the far end and an acknowledgment to return, in seconds. The raw result is in bits; dividing by 8 converts it to bytes, the more common unit for buffer and window sizing. For example, a 100 Mbps link with a 50 ms RTT has a BDP of 100,000,000 × 0.05 = 5,000,000 bits, or 625,000 bytes (about 610 KiB).
Why the TCP window matters
- TCP can only have as much unacknowledged data in flight as its window (or buffer) allows. If the window is smaller than the BDP, the sender must pause for acknowledgments before sending more, and throughput falls below the link's capacity even though the link itself is not saturated.
- Classic TCP's window field is 16 bits, capping the window at 65,535 bytes (64 KB) unless the TCP Window Scaling option (RFC 1323) is negotiated between both ends of the connection.
- Links where the BDP exceeds that 64 KB limit are commonly called "long fat networks" (LFNs) — high bandwidth, high latency, or both — and window scaling is required on them to reach full throughput.
Good-enough vs. precise
This calculator uses the standard BDP formula and a single flow's window size — it is a planning and diagnostic estimate, not a live network measurement. Real throughput is also shaped by packet loss, the congestion-control algorithm in use, competing traffic, and router buffering, so treat the result as the theoretical ceiling for one connection; measured throughput from a tool such as iperf will typically sit at or below this figure.