1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
|
<!DOCTYPE html>
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
<meta name="viewport" content="width=device-width,initial-scale=1,user-scalable=no" />
<!-- oktavia -->
<link rel="stylesheet" href="assets/searchstyle.css" type="text/css" />
<script src="search/jquery-1.9.1.min.js"></script>
<script src="search/oktavia-jquery-ui.js"></script>
<script src="search/oktavia-english-search.js"></script>
<!-- /oktavia -->
<link rel="stylesheet" href="assets/style.css" type="text/css" />
<title>Benchmarks - H2O - the optimized HTTP/2 server</title>
</head>
<body>
<div id="body">
<div id="top">
<h1>
<a href="index.html">H2O</a>
</h1>
<p class="description">the optimized HTTP/1.x, HTTP/2 server</p>
<!-- oktavia -->
<form id="searchform">
<input class="search" type="search" name="search" id="search" results="5" value="" placeholder="Search" />
<div id="searchresult_box">
<div id="close_search_box">×</div>
<div id="searchresult_summary"></div>
<div id="searchresult"></div>
<div id="searchresult_nav"></div>
<span class="pr">Powered by <a href="https://github.com/shibukawa/oktavia">Oktavia</a></span>
</div>
</form>
<!-- /oktavia -->
</div>
<table id="menu">
<tr>
<td><a href="index.html">Top</a></td>
<td><a href="install.html">Install</a></td>
<td><a href="configure.html">Configure</a></td>
<td><a href="faq.html">FAQ</a></td>
<td><a href="http://blog.kazuhooku.com/search/label/H2O" target="_blank">Blog</a></td>
<td><a href="http://github.com/h2o/h2o/" target="_blank">Source</a></td>
</tr>
</table>
<div id="main">
<h2>
Benchmarks
</h2>
<h3 id="download-timings">Download Timings</h3>
<div>
<p>
Providing quick response to user is more important than anything else in web performance tuning.
According to a research conducted by Microsoft, 500msec slowdown in Bing causes their revenue go down by 1.2%<sup><a href="#note_1" id="#cite_1" title="Velocity and the Bottom Line - O'Reilly Radar">1</sup></a></sup>.
</p>
<p>
The chart below compares the first-paint times and download completion times of different web browsers / HTTP servers on a simulated network of 8Mbps bandwidth with 100ms latency, which is typical for today's mobile networks<sup><a href="#note_2" id="#cite_2" title="A fork of http2rulez.com was used as the target website; bandwidth and latency were induced to local network using qdisc, specifically by running tc qdisc replace dev eth1 root handle 1:0 tbf rate 8192kbit burst 2048 latency 100ms; sudo tc qdisc add dev eth1 parent 1:1 netem delay 100ms, and sysctl -w net.ipv4.tcp_no_metrics_save=1.">2</sup></a></sup>.
</p>
<div align="center">
<a href="assets/8mbps100msec-nginx195-h2o150.png" target="_blank"><img src="assets/8mbps100msec-nginx195-h2o150.png" height="300"></a>
</div>
<p>
It is clear in the case of this benchmark that the visitors of the web site would be more satisfied, if H2O was used as the HTTP server.
</p>
</div>
<h3 id="static-file">Static-File Serving</h3>
<div>
<p>
Below chart shows the scores recorded on Amazon EC2 running two c3.8xlarge instances (server and client) on a single network placement, serving a 612-byte file<sup><a href="#note_3" id="#cite_3" title="Configuration files used: nginx.conf, h2o.conf.">3</sup></a></sup>.
For each measurement, 250 concurrent clients were used<sup><a href="#note_4" id="#cite_4" title="Wrk was used for HTTP/1 tests. h2load was used for HTTP/2.">4</sup></a></sup>.
<code>open_file_cache</code> was used for Nginx.
H2O implements a open-file-cache that gets updated immediately when the files are replaced.
</p>
<div align="center">
<a href="assets/staticfile612-nginx1910-h2o170.png" target="_blank"><img src="assets/staticfile612-nginx1910-h2o170.png" height="300"></a>
</div>
</div>
<h3 id="reverse-proxy">Reverse Proxy</h3>
<div>
<p>
Presented below is an old chart showing the scores recorded on Amazon EC2 running two c3.8xlarge instances (server and client) on a single network placement<sup><a href="#note_5" id="#cite_5" title="For reverse-proxy tests, another H2O process running on the same host was used as the upstream server">5</sup></a></sup><sup><a href="#note_6" id="#cite_6" title="open-file-cache was not used in the static-file benchmark">6</sup></a></sup>.
</p>
<div align="center">
<a href="assets/remotebench.png" target="_blank"><img src="assets/remotebench.png" width="400"></a>
</div>
</div>
<div class="notes">
<h3>Notes:</h3>
<ol>
<li id="note_1"><a href="http://radar.oreilly.com/2009/07/velocity-making-your-site-fast.html">Velocity and the Bottom Line - O'Reilly Radar</a></li>
<li id="note_2"><a href="https://github.com/kazuho/http2rulez.com">A fork of http2rulez.com</a> was used as the target website; bandwidth and latency were induced to local network using <a href="http://linux-ip.net/articles/Traffic-Control-HOWTO/components.html">qdisc</a>, specifically by running <code>tc qdisc replace dev eth1 root handle 1:0 tbf rate 8192kbit burst 2048 latency 100ms; sudo tc qdisc add dev eth1 parent 1:1 netem delay 100ms</code>, and <code>sysctl -w net.ipv4.tcp_no_metrics_save=1</code>.</li>
<li id="note_3">Configuration files used: <a href="https://gist.github.com/kazuho/def1e71281ed4ae07b95">nginx.conf</a>, <a href="https://gist.github.com/kazuho/969bb99bae31d67e01c4">h2o.conf</a>.</li>
<li id="note_4"><a href="https://github.com/wg/wrk">Wrk</a> was used for HTTP/1 tests. <a href="https://nghttp2.org/documentation/h2load-howto.html">h2load</a> was used for HTTP/2.</li>
<li id="note_5">For reverse-proxy tests, another H2O process running on the same host was used as the upstream server</li>
<li id="note_6">open-file-cache was not used in the static-file benchmark</li>
</ol>
</div>
</div>
<div id="footer">
<p>
Copyright © 2015 <a href="http://dena.com/intl/">DeNA Co., Ltd.</a> et al.
</p>
</div>
</body>
</html>
|
