<?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: StoneGate 5.0: HTTPS / SSL inspection</title>
	<atom:link href="http://stoneblog.stonesoft.com/2009/03/stonegate-50-https-ssl-inspection/feed/" rel="self" type="application/rss+xml" />
	<link>http://stoneblog.stonesoft.com/2009/03/stonegate-50-https-ssl-inspection/</link>
	<description>Share knowledge about StoneGate</description>
	<lastBuildDate>Wed, 08 Feb 2012 09:11:13 +0000</lastBuildDate>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
	<generator>http://wordpress.org/?v=3.3.1</generator>
	<item>
		<title>By: joona</title>
		<link>http://stoneblog.stonesoft.com/2009/03/stonegate-50-https-ssl-inspection/comment-page-1/#comment-81</link>
		<dc:creator>joona</dc:creator>
		<pubDate>Wed, 18 Mar 2009 08:04:51 +0000</pubDate>
		<guid isPermaLink="false">http://stoneblog.stonesoft.com/?p=633#comment-81</guid>
		<description>Hi,

For TheGhostInTheMachine at comment #4:

When I used the word &quot;terminate&quot; I actually meant that the firewall acts as the end-point for the SSL tunnel. I did not mean the traffic blocking / denying in this context.

We also support the traffic blocking (refuse -action in the access rules) that politely sends a TCP RST message back to the connection originator. No ICMP error messages are sent for the blocked TCP connections.

- Joona</description>
		<content:encoded><![CDATA[<p>Hi,</p>
<p>For TheGhostInTheMachine at comment #4:</p>
<p>When I used the word &#8220;terminate&#8221; I actually meant that the firewall acts as the end-point for the SSL tunnel. I did not mean the traffic blocking / denying in this context.</p>
<p>We also support the traffic blocking (refuse -action in the access rules) that politely sends a TCP RST message back to the connection originator. No ICMP error messages are sent for the blocked TCP connections.</p>
<p>- Joona</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: CaptainObvious</title>
		<link>http://stoneblog.stonesoft.com/2009/03/stonegate-50-https-ssl-inspection/comment-page-1/#comment-80</link>
		<dc:creator>CaptainObvious</dc:creator>
		<pubDate>Tue, 17 Mar 2009 10:57:24 +0000</pubDate>
		<guid isPermaLink="false">http://stoneblog.stonesoft.com/?p=633#comment-80</guid>
		<description>&lt;b&gt;Johan Lindgren&lt;/b&gt; said: &lt;i&gt;&quot;How configurable is the whitelist? I can whitelist http://www.foo.bar, but could I also whitelist *.foo.bar ? Could I even “whitelist” something like http://www.foo.bar/something/something/blah.html?
Can I be more lenient if a web site reputation is Super-Trustworthy-2000 ™ but uses a self signed certificate? &quot;&lt;/i&gt;

I&#039;d like to make a small clarification about the HTTPS Inspection Policy. It&#039;s not a whitelist in the sense of &quot;sites that we consider secure&quot; or &quot;certificates from trusted certificate authorities&quot;, but rather a list of domains that are excluded from decryption and inspection. My understanding is that the main reason for having the option for an HTTPS Inspection Policy has to do with possible legal restrictions: in some jurisdictions, decrypting and inspecting HTTPS traffic is prohibited by laws pertaining to the privacy of communications.

Additionally, you add the &lt;i&gt;domain&lt;/i&gt; name of the site(s) you want to exclude from decryption and inspection to the HTTPS Inspection Policy, not the URL. The list of domains is checked against the domain name in the server&#039;s certificate, so wildcards can only be used if they appear in the server&#039;s certificate.</description>
		<content:encoded><![CDATA[<p><b>Johan Lindgren</b> said: <i>&#8220;How configurable is the whitelist? I can whitelist <a href="http://www.foo.bar" rel="nofollow">http://www.foo.bar</a>, but could I also whitelist *.foo.bar ? Could I even “whitelist” something like <a href="http://www.foo.bar/something/something/blah.html?" rel="nofollow">http://www.foo.bar/something/something/blah.html?</a><br />
Can I be more lenient if a web site reputation is Super-Trustworthy-2000 ™ but uses a self signed certificate? &#8220;</i></p>
<p>I&#8217;d like to make a small clarification about the HTTPS Inspection Policy. It&#8217;s not a whitelist in the sense of &#8220;sites that we consider secure&#8221; or &#8220;certificates from trusted certificate authorities&#8221;, but rather a list of domains that are excluded from decryption and inspection. My understanding is that the main reason for having the option for an HTTPS Inspection Policy has to do with possible legal restrictions: in some jurisdictions, decrypting and inspecting HTTPS traffic is prohibited by laws pertaining to the privacy of communications.</p>
<p>Additionally, you add the <i>domain</i> name of the site(s) you want to exclude from decryption and inspection to the HTTPS Inspection Policy, not the URL. The list of domains is checked against the domain name in the server&#8217;s certificate, so wildcards can only be used if they appear in the server&#8217;s certificate.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: TheGhostInTheMachine</title>
		<link>http://stoneblog.stonesoft.com/2009/03/stonegate-50-https-ssl-inspection/comment-page-1/#comment-78</link>
		<dc:creator>TheGhostInTheMachine</dc:creator>
		<pubDate>Mon, 16 Mar 2009 16:10:38 +0000</pubDate>
		<guid isPermaLink="false">http://stoneblog.stonesoft.com/?p=633#comment-78</guid>
		<description>In addition, I see you state you support CRL, and I am very glad to see that (similar products from another vendor I have worked with in the past only allow me to use predefined CRLs, which is not ideal, in cases such as the recent fake SSL certificates that have been circulating many underground sites since the public announcement in Decmeber of the MD5 hash exploit allowing creation of certificates that appear to be legitimately signed by a CA using MD5 when in fact they were not.)

Can we define a generic CRL to check against? or specific things to consider distrusted (for example, I set my web browser to block any site claiming to require export-strength limited keys, as those will commonly be created during man-in-the-middle attacks, 40 bit encryption is quite easily brute forced for creating false certificates that match a valid hash.  I also don&#039;t want to have to try to secure a U.S. government office when I have, for example, the Microsoft certificate defaults, including foreign [potentionally hostile] government controlled servers.)

Finally, can we define which CAs to check with, for example, an intranet CA used for all our management network&#039;s encryption.</description>
		<content:encoded><![CDATA[<p>In addition, I see you state you support CRL, and I am very glad to see that (similar products from another vendor I have worked with in the past only allow me to use predefined CRLs, which is not ideal, in cases such as the recent fake SSL certificates that have been circulating many underground sites since the public announcement in Decmeber of the MD5 hash exploit allowing creation of certificates that appear to be legitimately signed by a CA using MD5 when in fact they were not.)</p>
<p>Can we define a generic CRL to check against? or specific things to consider distrusted (for example, I set my web browser to block any site claiming to require export-strength limited keys, as those will commonly be created during man-in-the-middle attacks, 40 bit encryption is quite easily brute forced for creating false certificates that match a valid hash.  I also don&#8217;t want to have to try to secure a U.S. government office when I have, for example, the Microsoft certificate defaults, including foreign [potentionally hostile] government controlled servers.)</p>
<p>Finally, can we define which CAs to check with, for example, an intranet CA used for all our management network&#8217;s encryption.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: TheGhostInTheMachine</title>
		<link>http://stoneblog.stonesoft.com/2009/03/stonegate-50-https-ssl-inspection/comment-page-1/#comment-77</link>
		<dc:creator>TheGhostInTheMachine</dc:creator>
		<pubDate>Mon, 16 Mar 2009 16:02:01 +0000</pubDate>
		<guid isPermaLink="false">http://stoneblog.stonesoft.com/?p=633#comment-77</guid>
		<description>When it terminates a connection, can you define what ICMP code to send with the termination, for example, a code-3 type-9 for administratively prohibited network (such as for Tor nodes) or a code-40 type4/5 for needing authorization/authentication to access a resource (in case of direct linking or session hijacking, such as is done with many of the network attacks used against e-commerce systems)</description>
		<content:encoded><![CDATA[<p>When it terminates a connection, can you define what ICMP code to send with the termination, for example, a code-3 type-9 for administratively prohibited network (such as for Tor nodes) or a code-40 type4/5 for needing authorization/authentication to access a resource (in case of direct linking or session hijacking, such as is done with many of the network attacks used against e-commerce systems)</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: joona</title>
		<link>http://stoneblog.stonesoft.com/2009/03/stonegate-50-https-ssl-inspection/comment-page-1/#comment-76</link>
		<dc:creator>joona</dc:creator>
		<pubDate>Mon, 16 Mar 2009 13:15:04 +0000</pubDate>
		<guid isPermaLink="false">http://stoneblog.stonesoft.com/?p=633#comment-76</guid>
		<description>Hi,

The HTTPS Inspection simply opens the SSL/TLS encryption and therefore allows the inspection of the HTTP content.

There is nothing new to the actual inspection part, beside of the decryption part. The same fingerprints / protocol analysis / browser version checks / etc that have been available before for the plain-text HTTP will now be available also for the HTTPS.

Thus this feature is not going to add any &quot;commercial blacklisting&quot; or other web site categorization functions.

And yes, the feature can be used to terminate the inbound HTTPS requests, inspect them and then forward to the web farm. I cannot say yet about the performance issues, though.</description>
		<content:encoded><![CDATA[<p>Hi,</p>
<p>The HTTPS Inspection simply opens the SSL/TLS encryption and therefore allows the inspection of the HTTP content.</p>
<p>There is nothing new to the actual inspection part, beside of the decryption part. The same fingerprints / protocol analysis / browser version checks / etc that have been available before for the plain-text HTTP will now be available also for the HTTPS.</p>
<p>Thus this feature is not going to add any &#8220;commercial blacklisting&#8221; or other web site categorization functions.</p>
<p>And yes, the feature can be used to terminate the inbound HTTPS requests, inspect them and then forward to the web farm. I cannot say yet about the performance issues, though.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Johan Lindgren</title>
		<link>http://stoneblog.stonesoft.com/2009/03/stonegate-50-https-ssl-inspection/comment-page-1/#comment-75</link>
		<dc:creator>Johan Lindgren</dc:creator>
		<pubDate>Fri, 13 Mar 2009 12:59:55 +0000</pubDate>
		<guid isPermaLink="false">http://stoneblog.stonesoft.com/?p=633#comment-75</guid>
		<description>Oooh, cool!

How does it know it&#039;s a malicious site or contains dangerous exploits/data/etc? 

- IDS &quot;Fingerprints&quot; should work just fine as the SSL now gets terminated and re-built before the potential victim?

- Peer review sites such as www.mywot.come? ( Yet another cool Finnish &quot;gadget&quot; btw) More &quot;open source&quot;/community oriented. 

- Sites that claim to be industry-leading reputation systems? I guess you could call them commercial blacklist sites. Not entirely sure how they actually work. You buy a subscription or something, or they&#039;re a part of some proxy package from a vendor.

How configurable is the whitelist? I can whitelist www.foo.bar, but could I also whitelist *.foo.bar ? Could I even &quot;whitelist&quot; something like www.foo.bar/something/something/blah.html?
Can I be more lenient if a web site reputation is Super-Trustworthy-2000 (tm) but uses a self signed certificate? That seal of trust was just a fantasy name for illustration purposes! :-)

Really looking forward to the launch. Really hope you&#039;ll get many opportunities to show how much better you are than the other big vendors. Really hope you&#039;ll be able to show-case it where ever appropriate. 

And by show-case, I&#039;m talking show what this baby can do! Not just hand out some brochures and whitepapers that doesn&#039;t really show how much more powerful Stonesoft management is in comparison with any of the other offerings currently available (read Checkpoint, Netscreen, etc ).

Keep these cool features coming!</description>
		<content:encoded><![CDATA[<p>Oooh, cool!</p>
<p>How does it know it&#8217;s a malicious site or contains dangerous exploits/data/etc? </p>
<p>- IDS &#8220;Fingerprints&#8221; should work just fine as the SSL now gets terminated and re-built before the potential victim?</p>
<p>- Peer review sites such as <a href="http://www.mywot.come?" rel="nofollow">http://www.mywot.come?</a> ( Yet another cool Finnish &#8220;gadget&#8221; btw) More &#8220;open source&#8221;/community oriented. </p>
<p>- Sites that claim to be industry-leading reputation systems? I guess you could call them commercial blacklist sites. Not entirely sure how they actually work. You buy a subscription or something, or they&#8217;re a part of some proxy package from a vendor.</p>
<p>How configurable is the whitelist? I can whitelist <a href="http://www.foo.bar" rel="nofollow">http://www.foo.bar</a>, but could I also whitelist *.foo.bar ? Could I even &#8220;whitelist&#8221; something like <a href="http://www.foo.bar/something/something/blah.html?" rel="nofollow">http://www.foo.bar/something/something/blah.html?</a><br />
Can I be more lenient if a web site reputation is Super-Trustworthy-2000 &#8482; but uses a self signed certificate? That seal of trust was just a fantasy name for illustration purposes! <img src='http://stoneblog.stonesoft.com/wp-includes/images/smilies/icon_smile.gif' alt=':-)' class='wp-smiley' /> </p>
<p>Really looking forward to the launch. Really hope you&#8217;ll get many opportunities to show how much better you are than the other big vendors. Really hope you&#8217;ll be able to show-case it where ever appropriate. </p>
<p>And by show-case, I&#8217;m talking show what this baby can do! Not just hand out some brochures and whitepapers that doesn&#8217;t really show how much more powerful Stonesoft management is in comparison with any of the other offerings currently available (read Checkpoint, Netscreen, etc ).</p>
<p>Keep these cool features coming!</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: dfoley</title>
		<link>http://stoneblog.stonesoft.com/2009/03/stonegate-50-https-ssl-inspection/comment-page-1/#comment-74</link>
		<dc:creator>dfoley</dc:creator>
		<pubDate>Fri, 13 Mar 2009 12:35:31 +0000</pubDate>
		<guid isPermaLink="false">http://stoneblog.stonesoft.com/?p=633#comment-74</guid>
		<description>Will this also work on inbound traffic - to protect a busy web farm with mant HTTPS servers ?  - What is the impact on Firewall utilization? 

I am looking forward to this capability and the other features of V5

DFoley</description>
		<content:encoded><![CDATA[<p>Will this also work on inbound traffic &#8211; to protect a busy web farm with mant HTTPS servers ?  &#8211; What is the impact on Firewall utilization? </p>
<p>I am looking forward to this capability and the other features of V5</p>
<p>DFoley</p>
]]></content:encoded>
	</item>
</channel>
</rss>

