<?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: High-Performance DNS for The Cloud</title>
	<atom:link href="http://www.igvita.com/2008/09/22/high-performance-dns-for-the-cloud/feed/" rel="self" type="application/rss+xml" />
	<link>http://www.igvita.com/2008/09/22/high-performance-dns-for-the-cloud/</link>
	<description>A goal is a dream with a deadline.</description>
	<pubDate>Thu, 11 Mar 2010 05:00:55 +0000</pubDate>
	<generator>http://wordpress.org/?v=2.7.1</generator>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
		<item>
		<title>By: Ilya Grigorik</title>
		<link>http://www.igvita.com/2008/09/22/high-performance-dns-for-the-cloud/comment-page-1/#comment-224860</link>
		<dc:creator>Ilya Grigorik</dc:creator>
		<pubDate>Wed, 02 Dec 2009 14:59:14 +0000</pubDate>
		<guid isPermaLink="false">http://www.igvita.com/?p=193#comment-224860</guid>
		<description>Michal, if you control the hardware, you should be able to just reassign the IP address to a different MAC at will, effectively achieving the same service as Elastic IP. Having said that, you're right, an automated heartbeat + failover is a good thing to have! HA Linux and others is probably what you're what looking for. Although, I'm sure there are hardware solutions for this as well...</description>
		<content:encoded><![CDATA[<p>Michal, if you control the hardware, you should be able to just reassign the IP address to a different MAC at will, effectively achieving the same service as Elastic IP. Having said that, you&#8217;re right, an automated heartbeat + failover is a good thing to have! HA Linux and others is probably what you&#8217;re what looking for. Although, I&#8217;m sure there are hardware solutions for this as well&#8230;</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Michal Frackowiak</title>
		<link>http://www.igvita.com/2008/09/22/high-performance-dns-for-the-cloud/comment-page-1/#comment-224859</link>
		<dc:creator>Michal Frackowiak</dc:creator>
		<pubDate>Wed, 02 Dec 2009 14:28:56 +0000</pubDate>
		<guid isPermaLink="false">http://www.igvita.com/?p=193#comment-224859</guid>
		<description>Yes, Elastic IPs are great, if you are using EC2 - you could set up a new server, even just to forward traffic.

However in conventional datacenters this is much more problematic - and we are facing this right now, having to design system that should distribute traffic to a dozen of equal servers. 

Anyway, the DNS balancer is a very good first-line solution, which can be supplemented by internal failover mechanisms, like EIP switching in EC2, or an equivalent mechanism in other datacenters or providers.</description>
		<content:encoded><![CDATA[<p>Yes, Elastic IPs are great, if you are using EC2 - you could set up a new server, even just to forward traffic.</p>
<p>However in conventional datacenters this is much more problematic - and we are facing this right now, having to design system that should distribute traffic to a dozen of equal servers. </p>
<p>Anyway, the DNS balancer is a very good first-line solution, which can be supplemented by internal failover mechanisms, like EIP switching in EC2, or an equivalent mechanism in other datacenters or providers.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Ilya Grigorik</title>
		<link>http://www.igvita.com/2008/09/22/high-performance-dns-for-the-cloud/comment-page-1/#comment-224858</link>
		<dc:creator>Ilya Grigorik</dc:creator>
		<pubDate>Wed, 02 Dec 2009 14:15:42 +0000</pubDate>
		<guid isPermaLink="false">http://www.igvita.com/?p=193#comment-224858</guid>
		<description>Michal, that's when services like Amazon's Elastic IP come in handy. Simply take that IP address (of the failed node) and reassign it to a different server. Takes ~dozens of seconds for the full switch on EC2.</description>
		<content:encoded><![CDATA[<p>Michal, that&#8217;s when services like Amazon&#8217;s Elastic IP come in handy. Simply take that IP address (of the failed node) and reassign it to a different server. Takes ~dozens of seconds for the full switch on EC2.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Michal Frackowiak</title>
		<link>http://www.igvita.com/2008/09/22/high-performance-dns-for-the-cloud/comment-page-1/#comment-224857</link>
		<dc:creator>Michal Frackowiak</dc:creator>
		<pubDate>Wed, 02 Dec 2009 13:56:06 +0000</pubDate>
		<guid isPermaLink="false">http://www.igvita.com/?p=193#comment-224857</guid>
		<description>DNS is indeed a cheap load balancing tool. But what if one of your frontend nodes fail? Good if you have low TTL, but if some ISPs cache your DNS records for 24 hours?

Any recommendations here?</description>
		<content:encoded><![CDATA[<p>DNS is indeed a cheap load balancing tool. But what if one of your frontend nodes fail? Good if you have low TTL, but if some ISPs cache your DNS records for 24 hours?</p>
<p>Any recommendations here?</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Ilya Grigorik</title>
		<link>http://www.igvita.com/2008/09/22/high-performance-dns-for-the-cloud/comment-page-1/#comment-134235</link>
		<dc:creator>Ilya Grigorik</dc:creator>
		<pubDate>Tue, 23 Sep 2008 12:26:33 +0000</pubDate>
		<guid isPermaLink="false">http://www.igvita.com/?p=193#comment-134235</guid>
		<description>Ah, that's a good one. If we need to keep databases in sync, we'll take the most recent snapshot, bring it up and set it up as a slave of the current production database. Then, when we flip the switch on DNS, we reverse the relationship between the databases. That way, we can still do a live rollback. 

Of course, if part of the upgrade are database schemas, then it gets a bit more complicated.</description>
		<content:encoded><![CDATA[<p>Ah, that&#8217;s a good one. If we need to keep databases in sync, we&#8217;ll take the most recent snapshot, bring it up and set it up as a slave of the current production database. Then, when we flip the switch on DNS, we reverse the relationship between the databases. That way, we can still do a live rollback. </p>
<p>Of course, if part of the upgrade are database schemas, then it gets a bit more complicated.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Mean Business &#187; Blog Archive &#187; links for 2008-09-23</title>
		<link>http://www.igvita.com/2008/09/22/high-performance-dns-for-the-cloud/comment-page-1/#comment-134200</link>
		<dc:creator>Mean Business &#187; Blog Archive &#187; links for 2008-09-23</dc:creator>
		<pubDate>Tue, 23 Sep 2008 09:00:41 +0000</pubDate>
		<guid isPermaLink="false">http://www.igvita.com/?p=193#comment-134200</guid>
		<description>[...] High-Performance DNS for The Cloud - igvita.com (tags: AWS cloud deployment architecture) [...]</description>
		<content:encoded><![CDATA[<p>[...] High-Performance DNS for The Cloud - igvita.com (tags: AWS cloud deployment architecture) [...]</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: W. Andrew Loe III</title>
		<link>http://www.igvita.com/2008/09/22/high-performance-dns-for-the-cloud/comment-page-1/#comment-134009</link>
		<dc:creator>W. Andrew Loe III</dc:creator>
		<pubDate>Mon, 22 Sep 2008 16:47:10 +0000</pubDate>
		<guid isPermaLink="false">http://www.igvita.com/?p=193#comment-134009</guid>
		<description>How do you deal with database drift during the transition? Do you just point your new infrastructure at your production database and migrate it, designing so that your changes are backwards compatible?</description>
		<content:encoded><![CDATA[<p>How do you deal with database drift during the transition? Do you just point your new infrastructure at your production database and migrate it, designing so that your changes are backwards compatible?</p>
]]></content:encoded>
	</item>
</channel>
</rss>
