Add TLS Dynamic Record Resizing to Nginx
From a424fefb0a638eb6d32756b5a0c471efc63e5384 Mon Sep 17 00:00:00 2001 From: Vlad Krasnov
Date: Sat, 9 Jan 2016 06:53:14 -0800
Subject: [PATCH] - Add TLS Dynamic Record Resizing
What we do now:
We use a static record size of 4K. This gives a good balance of latency and throughput.
Optimize latency:
By initialy sending small (1 TCP segment) sized records, we are able to avoid HoL blocking of the first byte. This means TTFB is sometime lower by a whole RTT.
Optimizing throughput:
By sending increasingly larger records later in the connection, when HoL is not a problem, we reduce the overhead of TLS record (29 bytes per record with GCM/CHACHA-POLY).
Logic:
Start each connection with small records (1369 byte default, change with ssl_dyn_rec_size_lo). After a given number of records (40, change with ssl_dyn_rec_threshold) start sending larger records (4229, ssl_dyn_rec_size_hi). Eventually after the same number of records, start sending the largest records (ssl_buffer_size).
In case the connection idles for a given amount of time (1s, ssl_dyn_rec_timeout), the process repeats itself (i.e. begin sending small records again).
Upstream source: