slow requests using node-http-proxy on ubuntu

node-http-proxy slow performance

var http = require('http');
var httpProxy = require('http-proxy');

//
// Proxy Server
//
var agent = new http.Agent();
agent.maxSockets = 10000;

var options = {
        agent: agent,
  	target:'http://localhost:9003'
};

var proxy = httpProxy.createProxyServer(options);

var proxyServer = http.createServer(function(req, res) {
  proxy.web(req, res, options);
});

proxyServer.on('connection', function (socket) {
    'use strict';
    socket.setNoDelay(true);
});

proxyServer.listen(8003);

//
// Target Http Server
//
var httpServer = http.createServer(function (req, res) {
  res.writeHead(200, { 'Content-Type': 'text/plain' });
  res.write('request successfully proxied to: ' + req.url + '\n' + JSON.stringify(req.headers, true, 2));
  res.end();
});

httpServer.on('connection', function (socket) {
    'use strict';
    socket.setNoDelay(true);
});

httpServer.listen(9003);

/*
client machine (running wrk from) AND server machine (running proxy.js):
- "Linux user1 3.11.0-12-generic #19-Ubuntu SMP Wed Oct 9 16:20:46 UTC 2013 x86_64 x86_64 x86_64 GNU/Linux, Ubuntu 13.10"
- Node v0.10.28

kernel-tuning done on both machines:
- (/etc/sysctl.conf):
  fs.file-max = 999999
  net.ipv4.tcp_rmem = 4096 4096 16777216
  net.ipv4.tcp_wmem = 4096 4096 16777216
  net.ipv4.ip_local_port_range = 1024 65535
- (/etc/security/limits.conf):
  #<domain>      <type>  <item>         <value>
  *               soft    nofile          10000
  *               hard    nofile          30000

benchmark with wrk (v3.1.0):

(direct):
wkr -c 100 -t 1 http://10.10.10.1:9003/ :

Running 10s test @ http://10.10.10.1:9003/
  1 threads and 100 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency    23.68ms    5.15ms  61.28ms   80.80%
    Req/Sec     4.30k   559.95     5.50k    73.33%
  41758 requests in 10.00s, 8.44MB read
Requests/sec:   4175.61
Transfer/sec:    864.48KB

(via proxy):
wkr -c 100 -t 1 http://10.10.10.1:8003/ :

Running 10s test @ http://10.10.10.1:8003/
  1 threads and 100 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency   275.81ms   20.41ms 331.75ms   74.06%
    Req/Sec   363.44     33.51   463.00     87.50%
  3614 requests in 10.00s, 854.09KB read
Requests/sec:    361.40
Transfer/sec:     85.41KB
*/