3proxy/bench

Benchmarks

Compares the speed and the cost of 3proxy versions, and of other proxies doing the same job, on Linux.

How the measurement is isolated

The client, the proxy and the origin each run in a network namespace of their own, linked by veth pairs:

[bcli] 10.0.1.2 --veth-- 10.0.1.1 [bprx] 10.0.2.1 --veth-- 10.0.2.2 [bsrv]
 wrk, iperf3              the proxy              nginx (files), iperf3 -s

so they share no loopback, and a packet crosses the same path whichever proxy is measured. Each role also runs in a cgroup of its own pinned to separate CPUs (--cpus, default cli=0-3,prx=4-5,srv=6-7), so one does not take CPU from another, and RPS steers what each veth receives to the receiver's CPUs. The proxy's cgroup is created anew for every run, which is what makes its CPU time and peak memory readable on their own.

Every proxy is given the same CPUs: nginx gets as many workers and haproxy as many threads as the proxy cgroup has CPUs; 3proxy runs a thread per connection inside the same cpuset. Logging is off everywhere.

none is no proxy at all: the client reaches the origin through the proxy namespace with the kernel forwarding, the upper bound for every scenario.

What is reported

For every run, in a CSV under bench/results/:

column	meaning
gbit, rate	throughput, requests per second for HTTP
p50_ms, p99_ms	latency from wrk (closed loop: compare at equal load only)
prx_cpu_s	CPU seconds the proxy used, all its threads and processes
cpu_s_per_gb, cpu_us_per_req	the proxy's cost per unit of work: the fairest figure, least affected by which side was the limit
peak_mb	the proxy's peak memory
cli_util, prx_util, srv_util	how busy each role's CPUs were: a client or origin near 100% means the result says nothing about the proxy
errors, retrans	socket errors and non-2xx answers; TCP retransmissions in the stream scenarios

Repetitions rotate the order of the proxies, so drift over time (heat, the host) falls on all of them alike; the summary gives medians.

Scenarios

name	load
stream1, stream1-down	one iperf3 stream through a TCP relay (3proxy tcppm, nginx stream, haproxy mode tcp), upload and download
stream16	sixteen streams
http-ka-1k, http-ka-1m	HTTP over kept-alive connections through an HTTP proxy (absolute URIs), 1 KB and 1 MB answers
http-close-1k	a new connection for every request

nginx is not a forward proxy: it answers absolute URIs by passing them to its upstream, which is the origin here, and it has no CONNECT or SOCKS, so only these scenarios compare like with like.

Running

Root, iproute2, iperf3, wrk, nginx (with the stream module: libnginx-mod-stream on Debian and Ubuntu) and, to compare with it, haproxy.

bench/build.sh 0.9.9.1 1.0.1                      # tags into bench/bin/
NAME=devel bench/build.sh .                       # the working tree
CC="gcc -fcommon" bench/build.sh 0.8.13           # before 0.9, newer gcc

sudo bench/bench.py \
    --proxy none \
    --proxy 3proxy:bench/bin/3proxy-1.0.1 \
    --proxy 3proxy:bench/bin/3proxy-devel \
    --proxy "3proxy:bench/bin/3proxy-0.9.4:-s0" \
    --proxy nginx --proxy haproxy

After the binary, a 3proxy entry can carry options for both its services: -s0 keeps 0.9.0 to 0.9.8, which use splice() by default, from using it, and -ocTCP_NODELAY -osTCP_NODELAY turns Nagle's algorithm off on the client and server sockets.

--scenario picks scenarios, --reps and --duration set the repetitions and their length, --netem "delay 1ms rate 10gbit" puts a slower, longer link between the client and the proxy, --mtu changes the links' MTU.

Caveats

• A virtual machine cannot pin its virtual CPUs to physical ones, so its absolute numbers move between sessions: compare within one session, and keep absolute figures for real hardware.
• wrk measures latency in a closed loop, so a slower proxy is also offered less load; compare latencies only between runs of similar throughput.
• Versions are built with the Makefile they shipped with, so compiler flags differ between them (0.8 builds with -O2, 0.9.9 and later with -O3 -flto) as they did for their users.