<?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: SSL Offloading</title>
	<atom:link href="http://mcwresearch.com/archives/447/feed" rel="self" type="application/rss+xml" />
	<link>http://mcwresearch.com/archives/447</link>
	<description>Things I think I've thought about</description>
	<lastBuildDate>Thu, 16 Feb 2012 16:52:27 +0000</lastBuildDate>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
	<generator>http://wordpress.org/?v=3.3.1</generator>
<xhtml:meta xmlns:xhtml="http://www.w3.org/1999/xhtml" name="robots" content="noindex" />
	<item>
		<title>By: LonerVamp</title>
		<link>http://mcwresearch.com/archives/447/comment-page-1#comment-1007</link>
		<dc:creator>LonerVamp</dc:creator>
		<pubDate>Fri, 30 Mar 2007 01:44:29 +0000</pubDate>
		<guid isPermaLink="false">http://mcwresearch.com/archives/447#comment-1007</guid>
		<description>I tend to agree with you. I think it might be important to make sure readers (this might be more for Farnum than your article) are aware which direction SSL decryption is dealing with: egress to external connections or ingress to web servers you own.

You can terminate SSL at the hardware device. In fact, that&#039;s the whole point of it. You&#039;re not only offloading the SSL decryption/encryption burden from the web server (typically the most intensive native actions on a web server) but you&#039;re also saving money in certificate licensing. One cert is oftimes needed per web server...or per hardware SSL terminator. So one load-balancer or SSL terminator for 10 web servers still just means 1 for-pay cert.

And, yes, while Alan has a point that you now have unencrypted communication between device and web servers, you can also encrypt that communication using your own internal self-signed certs if you want. In fact, that&#039;s what we do on our NLB (of course, we lose the gains we made on web server load, but we&#039;re ok with that).

You&#039;re absolutely correct. Even if that internal traffic is unencrypted from device to web servers, if someone is able to sniff that, you have other more important problems.

On a side note, even if you hack an SSL terminating device, I&#039;d be curious if anyone has yet been able to hijack that device in a way that will leak the unencrypted information inside the device (i.e. before it hits the wire on the internal side). I&#039;d be impressed if someone has done that... (And you&#039;re right to tackle that by saying it&#039;s not much different than hijacking the web server anyway...)</description>
		<content:encoded><![CDATA[<p>I tend to agree with you. I think it might be important to make sure readers (this might be more for Farnum than your article) are aware which direction SSL decryption is dealing with: egress to external connections or ingress to web servers you own.</p>
<p>You can terminate SSL at the hardware device. In fact, that&#8217;s the whole point of it. You&#8217;re not only offloading the SSL decryption/encryption burden from the web server (typically the most intensive native actions on a web server) but you&#8217;re also saving money in certificate licensing. One cert is oftimes needed per web server&#8230;or per hardware SSL terminator. So one load-balancer or SSL terminator for 10 web servers still just means 1 for-pay cert.</p>
<p>And, yes, while Alan has a point that you now have unencrypted communication between device and web servers, you can also encrypt that communication using your own internal self-signed certs if you want. In fact, that&#8217;s what we do on our NLB (of course, we lose the gains we made on web server load, but we&#8217;re ok with that).</p>
<p>You&#8217;re absolutely correct. Even if that internal traffic is unencrypted from device to web servers, if someone is able to sniff that, you have other more important problems.</p>
<p>On a side note, even if you hack an SSL terminating device, I&#8217;d be curious if anyone has yet been able to hijack that device in a way that will leak the unencrypted information inside the device (i.e. before it hits the wire on the internal side). I&#8217;d be impressed if someone has done that&#8230; (And you&#8217;re right to tackle that by saying it&#8217;s not much different than hijacking the web server anyway&#8230;)</p>
]]></content:encoded>
	</item>
</channel>
</rss>

