ZenossZenoss

blog

Amazon Outages Equals "Down Goes the Internet"? Are you Kidding Me?

Share on facebook
Share on twitter
Share on linkedin
Share on reddit
Share on pocket

Even as I begin to write this blog posting, I can feel my blood pressure rising and my anger percolating.  I am tired of the supposed Cloud computing “experts” who continue to regurgitate other people’s thoughts or make ridiculous statements in the name of attracting attention.  These individuals seek fame, re-tweets and page-views by inflicting a negativity that is the definition of tabloid/sensational blogging/journalism.

Normally, I don’t call out these individuals directly, but I’ve had enough.  On August 9, 2011, John Ribeiro, IDG News Service wrote about the Amazon EC2 outage saying, “The incident which lasted for about 30 minutes however highlights the risks involved in a cloud computing model.”

How does this highlight the risks involved with cloud computing?  How many private data centers were knocked out by this act of mother nature?  How long did others public/private take to respond to the outage?  In reality, the lightning strike serves as a reminder to all IT operations that variables beyond our control are lurking and it is our responsibility to have a disaster plan that maintain business continuity.

The brilliance of Amazon EC2 is it is a single-purpose cloud (Xen) that allows for multi-purpose applications.  Amazon is supplying the infrastructure and connectivity; customers are supplying the intellectual property.  Attacks on Amazon have become commonplace because they are the undisputed leader in this industry and are within the public domain.  Every incident and/or outage is viewable for the entire world to see.  What if your company was required to show their outage/incident statistics?  We will start with MTTR and MTTF and mix in the number of services affected.

How about MG Siegler’s article on Techcrunch.com entitled “Down Goes the Internet….Again.  Amazon EC2 Outage Takes Down Foursquare, Instagram, Quora, Reddit, etc.”

Excuse me; “Down Goes The Internet” are you kidding me?  Mr. Siegler even had the gall to write, “…pretty much all of Twitter, Amazon’s EC2 service appears to be down.”  Once again, without any hard evidence a ridiculous claim is made within the blogosphere.  The reality was, and he did correct it, that Amazon was having an issue with EAST-1 not the entire Amazon cloud.

Yes, Amazon had an issue but why aren’t we talking of the failings of the companies that utilize their services?  How can companies that provide services be left off the hook for outages by simply pointing the finger at Amazon?  Didn’t anyone learn from the outages of the past?

Simply put, any public or private Cloud or legacy infrastructure is going to go down.  I learned this lessen in the late 90’s while experiencing a nationwide frame relay outage at AT&T.  Heck, even United Airlines had had issues keeping their mainframe applications running that effected operations.  The key is to execute on a plan to maintain business continuity and minimize customer downtime.  I don’t care if Amazon, Microsoft, or your internal IT organization hosts your services, you must have a plan in place to keep your business running.

Whether you like it or not, the Cloud computing IT inflection point is here to stay.  Stop the sensationalism, and let’s get down to the business of providing customers the proper Cloud instrumentation to turn these outages into incidents!

Image Credit:  the real Kam75

Categories

Subscribe

Enter your email address in the box below to subscribe to our blog.

Loading
FEATURED CONTENT
Analyst Report
The Forrester Wave™: Intelligent Application and Service Monitoring, Q2 2019
Analyst Report
Gartner Market Guide for AIOps Platforms

Enabling IT to Move at the Speed of Business

Zenoss is built for modern IT infrastructures. Let's discuss how we can work together.

Schedule a Demo

Want to see us in action? Schedule a demo today.