NOTE
1.3 TCP Flow Control
How TCP's receiver-advertised window prevents a fast sender from overrunning the receiver, including sliding windows, zero windows, and window scaling.
This is a historical learning note and may contain outdated or incomplete understanding.
1. Goal
TCP flow control protects the receiver from a sender that would otherwise deliver data faster than the receiving application and buffers can consume it.
This is different from congestion control, which protects the network path.
2. Receiver-Advertised Window
The receiver advertises how much additional receive-buffer space it is prepared to accept. This value is commonly called rwnd.
As bytes are acknowledged and consumed by the receiving application, the usable window moves forward—hence the term sliding window.
A sender must keep unacknowledged in-flight data within the effective send window.
3. Zero Window
If the receiver advertises a zero window, ordinary new data transmission pauses because the receive buffer has no capacity.
TCP uses window-probe/persist behavior to avoid a permanent deadlock if the later window-update segment is lost.
4. Window Scaling
The original TCP header window field is 16 bits. Modern high-bandwidth-delay-product paths need larger windows, so the window-scale option can be negotiated during the handshake.
5. Relationship to Congestion Control
A simplified view of the sender’s allowed in-flight data is:
send limit ≈ min(rwnd, cwnd)rwnd: receiver capacity;cwnd: network-congestion estimate.
The smaller constraint wins.