<?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 for Load Balancing Digest</title>
	<atom:link href="http://lbdigest.com/comments/feed/" rel="self" type="application/rss+xml" />
	<link>http://lbdigest.com</link>
	<description>Server Load Balancing Articles and News</description>
	<pubDate>Tue, 06 Jan 2009 05:27:46 +0000</pubDate>
	<generator>http://wordpress.org/?v=2.7</generator>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
		<item>
		<title>Comment on When Your Load Balancer Has A Short Attention Span by Bill Kish</title>
		<link>http://lbdigest.com/2008/11/20/when-your-load-balancer-has-a-short-attention-span/comment-page-1/#comment-1030</link>
		<dc:creator>Bill Kish</dc:creator>
		<pubDate>Sat, 13 Dec 2008 13:18:56 +0000</pubDate>
		<guid isPermaLink="false">http://lbdigest.com/?p=186#comment-1030</guid>
		<description>Good article Tony,

This option is called "once only" on Coyote Point Equalizer appliances. On software versions prior to 8.5, once_only is the default for newly added clusters which means that only the first set of headers will be parsed (That's where the name comes from, only parse the first request in a stream). We've found hat having this default causes more trouble than it's worth so we've switched the default to parse every request in the stream.</description>
		<content:encoded><![CDATA[<p>Good article Tony,</p>
<p>This option is called &#8220;once only&#8221; on Coyote Point Equalizer appliances. On software versions prior to 8.5, once_only is the default for newly added clusters which means that only the first set of headers will be parsed (That&#8217;s where the name comes from, only parse the first request in a stream). We&#8217;ve found hat having this default causes more trouble than it&#8217;s worth so we&#8217;ve switched the default to parse every request in the stream.</p>
]]></content:encoded>
	</item>
	<item>
		<title>Comment on 3 Things You Need To Know About Etherchannel by perh</title>
		<link>http://lbdigest.com/2008/07/10/3-things-you-need-to-know-about-etherchannel/comment-page-1/#comment-1029</link>
		<dc:creator>perh</dc:creator>
		<pubDate>Thu, 04 Dec 2008 12:53:49 +0000</pubDate>
		<guid isPermaLink="false">http://lbdigest.com/?p=160#comment-1029</guid>
		<description>Thanks. Great explanation</description>
		<content:encoded><![CDATA[<p>Thanks. Great explanation</p>
]]></content:encoded>
	</item>
	<item>
		<title>Comment on When Your Load Balancer Has A Short Attention Span by Malcolm Turnbull</title>
		<link>http://lbdigest.com/2008/11/20/when-your-load-balancer-has-a-short-attention-span/comment-page-1/#comment-1025</link>
		<dc:creator>Malcolm Turnbull</dc:creator>
		<pubDate>Thu, 27 Nov 2008 20:23:52 +0000</pubDate>
		<guid isPermaLink="false">http://lbdigest.com/?p=186#comment-1025</guid>
		<description>For the Loadbalancer.org appliances we use the open source haproxy for layer 7 connection handling. By default we force keepalive off (haproxy setting not on the real servers) for all the servers and for Apache &#38; IIS we feel thats a pretty sensible way to go. Haproxy also has the option to increase the listen window. We think its great and high performance but lacks a large number of layer 7 features. As it is open source we hope to start adding these features as soon as possible.</description>
		<content:encoded><![CDATA[<p>For the Loadbalancer.org appliances we use the open source haproxy for layer 7 connection handling. By default we force keepalive off (haproxy setting not on the real servers) for all the servers and for Apache &amp; IIS we feel thats a pretty sensible way to go. Haproxy also has the option to increase the listen window. We think its great and high performance but lacks a large number of layer 7 features. As it is open source we hope to start adding these features as soon as possible.</p>
]]></content:encoded>
	</item>
	<item>
		<title>Comment on When Your Load Balancer Has A Short Attention Span by tony</title>
		<link>http://lbdigest.com/2008/11/20/when-your-load-balancer-has-a-short-attention-span/comment-page-1/#comment-1021</link>
		<dc:creator>tony</dc:creator>
		<pubDate>Fri, 21 Nov 2008 16:41:09 +0000</pubDate>
		<guid isPermaLink="false">http://lbdigest.com/?p=186#comment-1021</guid>
		<description>Hi Benjamin, 

Good question.  When your browser connects to a site, it's actually making lots of separate requests for the HTML, JPGs, etc.  But it's typically happening within the same TCP connection.  The source and destination IPs are the same, and don't get changed until you hit the load balancer. 

If the load balancer only looks at the first request, and sees it's a JPG, it will send every request contained within that TCP connection (JPG or not) to the JPG server farm.

In that case, you need the load balancer to look at every request.  It will then proxy those requests, and open up new TCP connections to the appropriate servers and forward them on (with the destination IP changed, and sometimes the source IP too).</description>
		<content:encoded><![CDATA[<p>Hi Benjamin, </p>
<p>Good question.  When your browser connects to a site, it&#8217;s actually making lots of separate requests for the HTML, JPGs, etc.  But it&#8217;s typically happening within the same TCP connection.  The source and destination IPs are the same, and don&#8217;t get changed until you hit the load balancer. </p>
<p>If the load balancer only looks at the first request, and sees it&#8217;s a JPG, it will send every request contained within that TCP connection (JPG or not) to the JPG server farm.</p>
<p>In that case, you need the load balancer to look at every request.  It will then proxy those requests, and open up new TCP connections to the appropriate servers and forward them on (with the destination IP changed, and sometimes the source IP too).</p>
]]></content:encoded>
	</item>
	<item>
		<title>Comment on When Your Load Balancer Has A Short Attention Span by Benjamin</title>
		<link>http://lbdigest.com/2008/11/20/when-your-load-balancer-has-a-short-attention-span/comment-page-1/#comment-1020</link>
		<dc:creator>Benjamin</dc:creator>
		<pubDate>Fri, 21 Nov 2008 14:50:23 +0000</pubDate>
		<guid isPermaLink="false">http://lbdigest.com/?p=186#comment-1020</guid>
		<description>Nice post, but I'm confused on one thing. You say:

"...trying to separate out traffic such as JPGs from HTML to send to different servers (web switching/Layer 7 switching), or if you’re trying to insert headers into every connection (such as the true source IP address, or an SSL header), then this is a big problem. "

But if I am putting JPGs and HTML on different servers, wouldn't they have different IP addresses, and therefore different TCP sessions, eliminating the issue?</description>
		<content:encoded><![CDATA[<p>Nice post, but I&#8217;m confused on one thing. You say:</p>
<p>&#8220;&#8230;trying to separate out traffic such as JPGs from HTML to send to different servers (web switching/Layer 7 switching), or if you’re trying to insert headers into every connection (such as the true source IP address, or an SSL header), then this is a big problem. &#8221;</p>
<p>But if I am putting JPGs and HTML on different servers, wouldn&#8217;t they have different IP addresses, and therefore different TCP sessions, eliminating the issue?</p>
]]></content:encoded>
	</item>
	<item>
		<title>Comment on Web Switching: What&#8217;s In A Name? by Load Balancing Digest &#187; Blog Archive &#187; When Your Load Balancer Has A Short Attention Span</title>
		<link>http://lbdigest.com/2008/05/12/web-switching-whats-in-a-name/comment-page-1/#comment-1019</link>
		<dc:creator>Load Balancing Digest &#187; Blog Archive &#187; When Your Load Balancer Has A Short Attention Span</dc:creator>
		<pubDate>Fri, 21 Nov 2008 01:03:37 +0000</pubDate>
		<guid isPermaLink="false">http://lbdigest.com/?p=125#comment-1019</guid>
		<description>[...] commonplace.  Most often it&#8217;s used in cookie -based persistence, but it&#8217;s also used in web switching, true-source IP resolution, and other tasks. [...]</description>
		<content:encoded><![CDATA[<p>[...] commonplace.  Most often it&#8217;s used in cookie -based persistence, but it&#8217;s also used in web switching, true-source IP resolution, and other tasks. [...]</p>
]]></content:encoded>
	</item>
	<item>
		<title>Comment on KEMP LoadMaster 1500 Pre-Release Review by Load Balancing Digest &#187; Blog Archive &#187; Web Application Firewall: What&#8217;s In A Name?</title>
		<link>http://lbdigest.com/2008/08/20/kemp-loadmaster-1500-pre-release-review/comment-page-1/#comment-1011</link>
		<dc:creator>Load Balancing Digest &#187; Blog Archive &#187; Web Application Firewall: What&#8217;s In A Name?</dc:creator>
		<pubDate>Wed, 12 Nov 2008 20:27:30 +0000</pubDate>
		<guid isPermaLink="false">http://lbdigest.com/?p=164#comment-1011</guid>
		<description>[...] my recent review of KEMP&#8217;s new LoadMaster software, which includes Web Application Firewall capabilities, Ofer Shezaf (from breach.com) had this to [...]</description>
		<content:encoded><![CDATA[<p>[...] my recent review of KEMP&#8217;s new LoadMaster software, which includes Web Application Firewall capabilities, Ofer Shezaf (from breach.com) had this to [...]</p>
]]></content:encoded>
	</item>
	<item>
		<title>Comment on HTTP Analyzers and ASICs by paja</title>
		<link>http://lbdigest.com/2008/10/29/http-analyzers-and-asics/comment-page-1/#comment-1009</link>
		<dc:creator>paja</dc:creator>
		<pubDate>Thu, 06 Nov 2008 22:12:58 +0000</pubDate>
		<guid isPermaLink="false">http://lbdigest.com/?p=183#comment-1009</guid>
		<description>Hi Tony,
just please post a note or link when You will find anything for Safari, I prefer HTTPFox, just because Lori^w DevCentral does :) and use FF for debugging, which is not much comfortable when FF is not default browser and IEx is a Fusion away...

-- 
paja</description>
		<content:encoded><![CDATA[<p>Hi Tony,<br />
just please post a note or link when You will find anything for Safari, I prefer HTTPFox, just because Lori^w DevCentral does <img src='http://lbdigest.com/wp-includes/images/smilies/icon_smile.gif' alt=':)' class='wp-smiley' /> and use FF for debugging, which is not much comfortable when FF is not default browser and IEx is a Fusion away&#8230;</p>
<p>&#8211;<br />
paja</p>
]]></content:encoded>
	</item>
	<item>
		<title>Comment on HTTP Analyzers and ASICs by Lori MacVittie</title>
		<link>http://lbdigest.com/2008/10/29/http-analyzers-and-asics/comment-page-1/#comment-1008</link>
		<dc:creator>Lori MacVittie</dc:creator>
		<pubDate>Wed, 29 Oct 2008 11:25:59 +0000</pubDate>
		<guid isPermaLink="false">http://lbdigest.com/?p=183#comment-1008</guid>
		<description>Tony, 

For Firefox there are a couple of other options that are killer: HTTPFox and Firebug. 

For HTTP analysis in general, I prefer HTTPFox. Straight forward, feels a lot like ethereal (sorry, Wireshark ;-)) and is pretty compact for a plug-in. 

Firebug offers most of the same options, but it's really aimed more at developers, IMO, and not so heavy on the network/protocol analysis side. 

Lori</description>
		<content:encoded><![CDATA[<p>Tony, </p>
<p>For Firefox there are a couple of other options that are killer: HTTPFox and Firebug. </p>
<p>For HTTP analysis in general, I prefer HTTPFox. Straight forward, feels a lot like ethereal (sorry, Wireshark ;-)) and is pretty compact for a plug-in. </p>
<p>Firebug offers most of the same options, but it&#8217;s really aimed more at developers, IMO, and not so heavy on the network/protocol analysis side. </p>
<p>Lori</p>
]]></content:encoded>
	</item>
	<item>
		<title>Comment on KEMP LoadMaster 1500 Pre-Release Review by Ofer Shezaf</title>
		<link>http://lbdigest.com/2008/08/20/kemp-loadmaster-1500-pre-release-review/comment-page-1/#comment-1006</link>
		<dc:creator>Ofer Shezaf</dc:creator>
		<pubDate>Wed, 29 Oct 2008 01:10:12 +0000</pubDate>
		<guid isPermaLink="false">http://lbdigest.com/?p=164#comment-1006</guid>
		<description>Systems supporting only snort rules and lacking a positive security model are not usually not considered a web application firewall, but rather an intrusion prevention system.</description>
		<content:encoded><![CDATA[<p>Systems supporting only snort rules and lacking a positive security model are not usually not considered a web application firewall, but rather an intrusion prevention system.</p>
]]></content:encoded>
	</item>
</channel>
</rss>
