<?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"
	>
<channel>
	<title>Comments on: Ongoing Amazon EC2 Observations</title>
	<atom:link href="http://feedblendr.com/blog/2007/08/21/ongoing-amazon-ec2-observations/feed/" rel="self" type="application/rss+xml" />
	<link>http://feedblendr.com/blog/2007/08/21/ongoing-amazon-ec2-observations/</link>
	<description>Blending you a delicious feed smoothie!</description>
	<pubDate>Thu, 04 Dec 2008 22:41:43 +0000</pubDate>
	<generator>http://wordpress.org/?v=2.5.1</generator>
		<item>
		<title>By: Gil Megidish</title>
		<link>http://feedblendr.com/blog/2007/08/21/ongoing-amazon-ec2-observations/#comment-28399</link>
		<dc:creator>Gil Megidish</dc:creator>
		<pubDate>Sun, 02 Mar 2008 06:16:43 +0000</pubDate>
		<guid isPermaLink="false">http://www.feedblendr.com/blog/2007/08/21/ongoing-amazon-ec2-observations/#comment-28399</guid>
		<description>Four more reasons AGAINST hosting from Amazon EC2:

1. No fixed IP (yet). This is too scary for me. Everybody serving long enough had their servers rebooted. Now think what happens when you need to modify dns records and wait 24-48 until all dns servers update.

2. No phone number or direct email. Just grab a look at the forums. It's really sad. People list their ami id, and instance-id and ask amazon guys to please-please-please check it now. I wouldn't mind to pay extra $$$ for a phone support.

3. Virtual environment is not suitable for everybody. I spent $250 for EC2 uptime (several servers) while moving one of my website into a completely load-balanced and distributed environment. Of course, this doesn't include the countless hours I spent rewriting code. At the end of everything, when I launch it, it blew in my face. I'm using a custom server which uses unix pipes to send data between two processes. Hmf! EC2's Xen is capped at like 10,000 context-switches a sec! :(    So stuff that used to take 0.5 sec on dedicated, now take 15 sec (which is considered timeout on my javascript side). I really get a nice distributed environment, but at 1/10 of the processing capacity :)

4. Compare prices. You can get far better deals on vps from other companies. Amazon has a minimum price of $72 (0.10 * 24 * 30) per month, without single use of bandwidth. Companies offer large traffic packages bundled in price. 500gb worth of IN traffic is worth a lot of extra money. For that price you can already purchase a dedicated server at godaddy (which has really good 24/7 phone support.)

But that's just my opinion. :o)
- gil</description>
		<content:encoded><![CDATA[<p>Four more reasons AGAINST hosting from Amazon EC2:</p>
<p>1. No fixed IP (yet). This is too scary for me. Everybody serving long enough had their servers rebooted. Now think what happens when you need to modify dns records and wait 24-48 until all dns servers update.</p>
<p>2. No phone number or direct email. Just grab a look at the forums. It&#8217;s really sad. People list their ami id, and instance-id and ask amazon guys to please-please-please check it now. I wouldn&#8217;t mind to pay extra $$$ for a phone support.</p>
<p>3. Virtual environment is not suitable for everybody. I spent $250 for EC2 uptime (several servers) while moving one of my website into a completely load-balanced and distributed environment. Of course, this doesn&#8217;t include the countless hours I spent rewriting code. At the end of everything, when I launch it, it blew in my face. I&#8217;m using a custom server which uses unix pipes to send data between two processes. Hmf! EC2&#8217;s Xen is capped at like 10,000 context-switches a sec! <img src='http://feedblendr.com/blog/wp-includes/images/smilies/icon_sad.gif' alt=':(' class='wp-smiley' />    So stuff that used to take 0.5 sec on dedicated, now take 15 sec (which is considered timeout on my javascript side). I really get a nice distributed environment, but at 1/10 of the processing capacity <img src='http://feedblendr.com/blog/wp-includes/images/smilies/icon_smile.gif' alt=':)' class='wp-smiley' /> </p>
<p>4. Compare prices. You can get far better deals on vps from other companies. Amazon has a minimum price of $72 (0.10 * 24 * 30) per month, without single use of bandwidth. Companies offer large traffic packages bundled in price. 500gb worth of IN traffic is worth a lot of extra money. For that price you can already purchase a dedicated server at godaddy (which has really good 24/7 phone support.)</p>
<p>But that&#8217;s just my opinion. :o)<br />
- gil</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Stelios</title>
		<link>http://feedblendr.com/blog/2007/08/21/ongoing-amazon-ec2-observations/#comment-17598</link>
		<dc:creator>Stelios</dc:creator>
		<pubDate>Mon, 15 Oct 2007 13:27:31 +0000</pubDate>
		<guid isPermaLink="false">http://www.feedblendr.com/blog/2007/08/21/ongoing-amazon-ec2-observations/#comment-17598</guid>
		<description>SLA is there now. So one prob is going out of this list.</description>
		<content:encoded><![CDATA[<p>SLA is there now. So one prob is going out of this list.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Bert Armijo</title>
		<link>http://feedblendr.com/blog/2007/08/21/ongoing-amazon-ec2-observations/#comment-13126</link>
		<dc:creator>Bert Armijo</dc:creator>
		<pubDate>Fri, 24 Aug 2007 17:35:41 +0000</pubDate>
		<guid isPermaLink="false">http://www.feedblendr.com/blog/2007/08/21/ongoing-amazon-ec2-observations/#comment-13126</guid>
		<description>Thanks for sharing your experiences. Service like EC2 and ours at 3tera are opening up new avenues for developers and having users like yourself that will share openly as you've done helps everyone to build better services.</description>
		<content:encoded><![CDATA[<p>Thanks for sharing your experiences. Service like EC2 and ours at 3tera are opening up new avenues for developers and having users like yourself that will share openly as you&#8217;ve done helps everyone to build better services.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: beau</title>
		<link>http://feedblendr.com/blog/2007/08/21/ongoing-amazon-ec2-observations/#comment-13049</link>
		<dc:creator>beau</dc:creator>
		<pubDate>Thu, 23 Aug 2007 17:31:52 +0000</pubDate>
		<guid isPermaLink="false">http://www.feedblendr.com/blog/2007/08/21/ongoing-amazon-ec2-observations/#comment-13049</guid>
		<description>khang: As you mention, not having an SLA of any kind is a concern for reliability. I'm working on some "external" monitoring for instances, because basically I think the assumptions here need to be 1. You're on your own for keeping track of connectivity, 2. Things will fail. Another option for all of this is that you could look into something like &lt;a href="http://weoceo.weogeo.com"&gt;WeoCEO&lt;/a&gt; which provides all sorts of monitoring/automated scaling features if that's pricing-appropriate for you.

fthiery: Thanks for the pointer - someone else actually mentioned rPath to me yesterday, so I'll have to check them out.</description>
		<content:encoded><![CDATA[<p>khang: As you mention, not having an SLA of any kind is a concern for reliability. I&#8217;m working on some &#8220;external&#8221; monitoring for instances, because basically I think the assumptions here need to be 1. You&#8217;re on your own for keeping track of connectivity, 2. Things will fail. Another option for all of this is that you could look into something like <a href="http://weoceo.weogeo.com">WeoCEO</a> which provides all sorts of monitoring/automated scaling features if that&#8217;s pricing-appropriate for you.</p>
<p>fthiery: Thanks for the pointer - someone else actually mentioned rPath to me yesterday, so I&#8217;ll have to check them out.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: fthiery</title>
		<link>http://feedblendr.com/blog/2007/08/21/ongoing-amazon-ec2-observations/#comment-13022</link>
		<dc:creator>fthiery</dc:creator>
		<pubDate>Thu, 23 Aug 2007 10:28:36 +0000</pubDate>
		<guid isPermaLink="false">http://www.feedblendr.com/blog/2007/08/21/ongoing-amazon-ec2-observations/#comment-13022</guid>
		<description>You should check rpath Linux (http://www.rpath.com/corp/) which offers rollback functionnality and automatic AMI generation.

Cheers,</description>
		<content:encoded><![CDATA[<p>You should check rpath Linux (http://www.rpath.com/corp/) which offers rollback functionnality and automatic AMI generation.</p>
<p>Cheers,</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: khang</title>
		<link>http://feedblendr.com/blog/2007/08/21/ongoing-amazon-ec2-observations/#comment-12910</link>
		<dc:creator>khang</dc:creator>
		<pubDate>Tue, 21 Aug 2007 23:03:09 +0000</pubDate>
		<guid isPermaLink="false">http://www.feedblendr.com/blog/2007/08/21/ongoing-amazon-ec2-observations/#comment-12910</guid>
		<description>hey, man.  thanks for sharing your experiences.  great stuff.

given the lack of an SLA, what's your take on the "reliability" of EC2?  what kinda things/precautions do you put in place to live with any reliability problems?

i just started my site about a week ago on one instance and things were going well until today.  i woke up to the first hiccup with my instance: for some reason, it decided to reboot itself, which freaked me out because i hadn't yet bound my changes to an AMI.</description>
		<content:encoded><![CDATA[<p>hey, man.  thanks for sharing your experiences.  great stuff.</p>
<p>given the lack of an SLA, what&#8217;s your take on the &#8220;reliability&#8221; of EC2?  what kinda things/precautions do you put in place to live with any reliability problems?</p>
<p>i just started my site about a week ago on one instance and things were going well until today.  i woke up to the first hiccup with my instance: for some reason, it decided to reboot itself, which freaked me out because i hadn&#8217;t yet bound my changes to an AMI.</p>
]]></content:encoded>
	</item>
</channel>
</rss>
