What’s the distinction between tcp movement management and http2 movement management.
Extra particularly, with window_update frames that work on the entire connection vs tcp window dimension.
To my understanding, window_update frames that work on the entire connection, works by a sender being allowed to ship as many octets as agreed upon by the preliminary window dimension within the settings body, the receiver then sends a window replace body to inform the sender their buffer is free and the sender can once more ship octets equal to the preliminary window dimension decided by the settings body.
with tcp movement management, every ACK by a receiver incorporates a window dimension subject that tells the sender what number of remaining octets the sender can ship.
This appears redundant to me. With TCP you’re telling the sender the up to date window dimension after each packet will get acquired and with http2 you wait till all the octets are acquired earlier than sending a window_update body saying your buffer is free and to ship octets equal to that preliminary window dimension once more.
Why the necessity for these two movement controls? Why couldnt http2 simply depend on tcp’s movement management?