<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	xmlns:sy="http://purl.org/rss/1.0/modules/syndication/"
		>
<channel>
	<title>Comments on: Sharding Dominant Domains</title>
	<atom:link href="http://www.stevesouders.com/blog/2009/05/12/sharding-dominant-domains/feed/" rel="self" type="application/rss+xml" />
	<link>http://www.stevesouders.com/blog/2009/05/12/sharding-dominant-domains/</link>
	<description>Essential knowledge for making your web pages faster.</description>
	<lastBuildDate>Sat, 04 Feb 2012 09:46:28 +0000</lastBuildDate>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
	<generator>http://wordpress.org/?v=3.2.1</generator>
	<item>
		<title>By: Brendan Gibson</title>
		<link>http://www.stevesouders.com/blog/2009/05/12/sharding-dominant-domains/#comment-3079</link>
		<dc:creator>Brendan Gibson</dc:creator>
		<pubDate>Fri, 04 Mar 2011 21:34:47 +0000</pubDate>
		<guid isPermaLink="false">http://www.stevesouders.com/blog/?p=370#comment-3079</guid>
		<description>Thanks Steve.  I&#039;ll see if we can do some of our own tests on this and feed the results back into the community.</description>
		<content:encoded><![CDATA[<p>Thanks Steve.  I&#8217;ll see if we can do some of our own tests on this and feed the results back into the community.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Steve Souders</title>
		<link>http://www.stevesouders.com/blog/2009/05/12/sharding-dominant-domains/#comment-3070</link>
		<dc:creator>Steve Souders</dc:creator>
		<pubDate>Fri, 04 Mar 2011 02:21:06 +0000</pubDate>
		<guid isPermaLink="false">http://www.stevesouders.com/blog/?p=370#comment-3070</guid>
		<description>@Brendan: Since newer browsers open &lt;b&gt;more&lt;/b&gt; connections per domain, it&#039;s probably better to revise the number downwards. I think 2 is a good compromize, but that&#039;s just a hunch. It&#039;d be great if some production property ran a test to determine the optimal number.</description>
		<content:encoded><![CDATA[<p>@Brendan: Since newer browsers open <b>more</b> connections per domain, it&#8217;s probably better to revise the number downwards. I think 2 is a good compromize, but that&#8217;s just a hunch. It&#8217;d be great if some production property ran a test to determine the optimal number.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Brendan Gibson</title>
		<link>http://www.stevesouders.com/blog/2009/05/12/sharding-dominant-domains/#comment-3045</link>
		<dc:creator>Brendan Gibson</dc:creator>
		<pubDate>Thu, 17 Feb 2011 17:57:57 +0000</pubDate>
		<guid isPermaLink="false">http://www.stevesouders.com/blog/?p=370#comment-3045</guid>
		<description>This article was written two years ago when IE7 may have been the dominant browser.  Now that the proportions of browser usage has changed, do you think we need to revisit the optimal number of domains?  And revise it upwards?</description>
		<content:encoded><![CDATA[<p>This article was written two years ago when IE7 may have been the dominant browser.  Now that the proportions of browser usage has changed, do you think we need to revisit the optimal number of domains?  And revise it upwards?</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Steve Souders</title>
		<link>http://www.stevesouders.com/blog/2009/05/12/sharding-dominant-domains/#comment-2015</link>
		<dc:creator>Steve Souders</dc:creator>
		<pubDate>Fri, 30 Apr 2010 13:57:56 +0000</pubDate>
		<guid isPermaLink="false">http://www.stevesouders.com/blog/?p=370#comment-2015</guid>
		<description>@Guillermo: The best test is to do a test on your live traffic with JavaScript timing the page (see the Jiffy Web project).

@Matt: You can look at the source code and see that these are the correct settings.

@Michael: I have not tested ports. It would be great if you created a test page with a bunch of images from different ports, and then watched them in Firebug Net Panel to see if the number of parallel connections exceeds the max. Please comment back here if you do so.</description>
		<content:encoded><![CDATA[<p>@Guillermo: The best test is to do a test on your live traffic with JavaScript timing the page (see the Jiffy Web project).</p>
<p>@Matt: You can look at the source code and see that these are the correct settings.</p>
<p>@Michael: I have not tested ports. It would be great if you created a test page with a bunch of images from different ports, and then watched them in Firebug Net Panel to see if the number of parallel connections exceeds the max. Please comment back here if you do so.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Michael</title>
		<link>http://www.stevesouders.com/blog/2009/05/12/sharding-dominant-domains/#comment-2014</link>
		<dc:creator>Michael</dc:creator>
		<pubDate>Fri, 30 Apr 2010 09:05:24 +0000</pubDate>
		<guid isPermaLink="false">http://www.stevesouders.com/blog/?p=370#comment-2014</guid>
		<description>Will domain sharding work if I serve images from different TCP ports?</description>
		<content:encoded><![CDATA[<p>Will domain sharding work if I serve images from different TCP ports?</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Matt</title>
		<link>http://www.stevesouders.com/blog/2009/05/12/sharding-dominant-domains/#comment-919</link>
		<dc:creator>Matt</dc:creator>
		<pubDate>Tue, 13 Oct 2009 15:23:52 +0000</pubDate>
		<guid isPermaLink="false">http://www.stevesouders.com/blog/?p=370#comment-919</guid>
		<description>&quot;Not all browsers are restricted to just two parallel downloads per hostname.&quot;

How sure are you about that?  My impression was that installing IE8 raises the simultaneous connections limit by changing a system-wide setting.

Have you tested Firefox or Chrome on a fresh install of Windows that&#039;s never seen IE8?</description>
		<content:encoded><![CDATA[<p>&#8220;Not all browsers are restricted to just two parallel downloads per hostname.&#8221;</p>
<p>How sure are you about that?  My impression was that installing IE8 raises the simultaneous connections limit by changing a system-wide setting.</p>
<p>Have you tested Firefox or Chrome on a fresh install of Windows that&#8217;s never seen IE8?</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Guillermo Quintana</title>
		<link>http://www.stevesouders.com/blog/2009/05/12/sharding-dominant-domains/#comment-607</link>
		<dc:creator>Guillermo Quintana</dc:creator>
		<pubDate>Fri, 22 May 2009 05:28:34 +0000</pubDate>
		<guid isPermaLink="false">http://www.stevesouders.com/blog/?p=370#comment-607</guid>
		<description>I&#039;ve been doing quite a bit of research regarding domain shredding. Is there a tool or method i could use ( short of proto-typing) to test using 2 vs. 4 host manes for static images across my website?  My testing is limited to only http requests as i would prefer to leave all of my assets for ssl under a single hostname</description>
		<content:encoded><![CDATA[<p>I&#8217;ve been doing quite a bit of research regarding domain shredding. Is there a tool or method i could use ( short of proto-typing) to test using 2 vs. 4 host manes for static images across my website?  My testing is limited to only http requests as i would prefer to leave all of my assets for ssl under a single hostname</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Éric</title>
		<link>http://www.stevesouders.com/blog/2009/05/12/sharding-dominant-domains/#comment-595</link>
		<dc:creator>Éric</dc:creator>
		<pubDate>Tue, 19 May 2009 10:37:15 +0000</pubDate>
		<guid isPermaLink="false">http://www.stevesouders.com/blog/?p=370#comment-595</guid>
		<description>Also, for SSL the cost of a new connection is higher, so having to much domain / to much connections could be negative far more quicker than with non-ssl domains.

With new browers that have 6 threads per domain, it&#039;s probably safe to stay with only one domain on SSL (and never more than two).</description>
		<content:encoded><![CDATA[<p>Also, for SSL the cost of a new connection is higher, so having to much domain / to much connections could be negative far more quicker than with non-ssl domains.</p>
<p>With new browers that have 6 threads per domain, it&#8217;s probably safe to stay with only one domain on SSL (and never more than two).</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Party Ark</title>
		<link>http://www.stevesouders.com/blog/2009/05/12/sharding-dominant-domains/#comment-592</link>
		<dc:creator>Party Ark</dc:creator>
		<pubDate>Mon, 18 May 2009 19:48:15 +0000</pubDate>
		<guid isPermaLink="false">http://www.stevesouders.com/blog/?p=370#comment-592</guid>
		<description>Good article - on thing to remember, however, is NOT to implement this technique for SSL (https) unless you&#039;re able to secure ALL the &#039;sharded&#039; subdomains. There are certificates that can handle multiple subdomains, but most don&#039;t. And if you don&#039;t have one, your user will end up with all sorts of warning messages about mixed resources and a broken &#039;padlock&#039; symbol.</description>
		<content:encoded><![CDATA[<p>Good article &#8211; on thing to remember, however, is NOT to implement this technique for SSL (https) unless you&#8217;re able to secure ALL the &#8216;sharded&#8217; subdomains. There are certificates that can handle multiple subdomains, but most don&#8217;t. And if you don&#8217;t have one, your user will end up with all sorts of warning messages about mixed resources and a broken &#8216;padlock&#8217; symbol.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Martin Borthiry</title>
		<link>http://www.stevesouders.com/blog/2009/05/12/sharding-dominant-domains/#comment-586</link>
		<dc:creator>Martin Borthiry</dc:creator>
		<pubDate>Thu, 14 May 2009 14:09:08 +0000</pubDate>
		<guid isPermaLink="false">http://www.stevesouders.com/blog/?p=370#comment-586</guid>
		<description>Great article!

 I think that in some cases could be possible to check the browser type and define dynamically the number of hostnames. The script has a time cost, but it reduce the DNS lookup time for newest browsers.</description>
		<content:encoded><![CDATA[<p>Great article!</p>
<p> I think that in some cases could be possible to check the browser type and define dynamically the number of hostnames. The script has a time cost, but it reduce the DNS lookup time for newest browsers.</p>
]]></content:encoded>
	</item>
</channel>
</rss>

