tag:blogger.com,1999:blog-4512988492228683846.post2405929778874191431..comments2023-08-28T02:29:11.306-07:00Comments on Rip Van Webble: Quic: the Elephant in the RoomMichael Thomashttp://www.blogger.com/profile/12086058018576918310noreply@blogger.comBlogger1125tag:blogger.com,1999:blog-4512988492228683846.post-7481673891086074872021-04-06T03:41:51.789-07:002021-04-06T03:41:51.789-07:00If I understand the protocol correctly, the goal a...If I understand the protocol correctly, the goal as you mention up top is latency reduction. And packet size doesn't actually correlate much with latency here. The assumption is that the transit time is far longer than the encode/decode time for any packets that are sent, and it also assumes that packets of different sizes travel at the same speed. If both those assumptions hold, then the PEM files could be ten times bigger with only a few nanoseconds' increase in latency for the protocol. And both seem to hold on the modern web.Phlosioneerhttps://www.blogger.com/profile/08351849425590687027noreply@blogger.com