Re: [quicwg/base-drafts] Byte counting at Congestion avoidance (#3917)

[Martin Thomson <[e-mail]>]

@martinthomson commented on this pull request.



> +      bytes_acked += acked_packet.sent_bytes
+      if (bytes_acked >= congestion_window):
+        bytes_acked -= congestion_window
+        congestion_window += max_datagram_size

Oh, I should probably concede that more increases to the congestion window will result in the final congestion window being smaller, so you are trading a larger final increase (this PR) with earlier availability of bytes (status quo).

Say you get 10 acks for 1000 bytes, with a starting cwnd of 10,000 and max_datagram_sizeo of 1000.

With this code, you end up with no change until the last ack, where you go to 11,000.

With the previous code, you would go to: 10100, 10199, 10297, 10394, 10490, 10585, 10680, 10773, 10866, 10958.  (That's using double-precision logic, you lose two bytes if you round.)

That's less in the end, but the value of that to you depends on whether you can use any of those intermediate values.

-- 
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/quicwg/base-drafts/pull/3917#discussion_r461339606