Re: Fwd: I-D Action: draft-melnikov-rfc2088bis-01.txt
Arnt Gulbrandsen <[email protected]>
| Newsgroups | gmane.ietf.imapext |
|---|---|
| Message-ID | <[email protected]> |
Alexey Melnikov writes: > I've seen lots of small messages around 3-4Kb, so how about > 4096 instead? It is still going to be a single TCP packet in > most cases. No, not one packet. What you mean is the typical TCP window, ie. the number of bytes a TCP sender can send without waiting for Ack from the peer. Which is often 4500 until the peers negotiate different values. I quite agree that 2048 is too low. Any value between 4096 and 32768 is okay, I think, and practically equivalent. Most messages are either <4096 (when leaving the sending MUA) or >32768. Arnt _______________________________________________ imapext mailing list [email protected] https://www.ietf.org/mailman/listinfo/imapext