In the wake of Google’s downtime we have decided at 48Web to develop a Cloud Disaster Recovery Plan for our business. Once completed we’ll be providing it to anyone interested in using it to model their own disaster recovery plan for “the cloud”.
What’s a disaster recovery plan for? In the event something bad happens – ie: servers go down – a disaster recovery plan is a set of processes that you have at hand to address the situation without hesitation. Generally these are reserved for larger companies or data centers. But as a small business whose entire IT infrastructure lives in the cloud – we thought it necessary to have a plan in place.
Here are the problems at hand and potential solutions. We want to use this blog post to crowd source ideas and conversations around cloud disaster recovery to better develop our outline and plan. Please, chime in!
Communications
Our entire communications infrastructure is in the cloud. We use Skype and Google Talk for voice communications. We use Google Apps to host our email. Down the list we also use Twitter, FriendFeed, and our blog for communication to clients and prospects.
If Google Apps goes down (as it did last week) and we are unable to receive email our new process is to write a blog post alerting our community that we are unable to receive email. This will be sent out to our email newsletter subscribers and customers (via aWeber). We’ll also send out a tweet from our personal and company accounts. We’ll be using our pre-defined content outposts and distribution system to communicate the message.
Storage
At 48Web we store all of our documents, media, etc on Amazon S3 and Dropbox. We also use Amazon’s content delivery network (CDN) CloudFront to store and serve all of our web content – from the internet business podcast to CSS files to images.
If Amazon S3 goes down – and it has – we are in trouble. Images, CSS, Javascripts won’t load. Our podcast won’t be available for download. In short – most of our multimedia will be inaccessible. Our current plan is to mirror our Amazon S3 files on Rackspace Cloud Files. In the event of Amazon S3 downtime we’ll be able to switch out URL’s for our most requested items to the Cloud Files URL.
What do you think? Have a better idea?
Web Hosting
We use Mosso for production hosting and Dreamhost to host our development / testing environments. Although we’ve yet to have extended downtime with Mosso there is still a chance. Our new process is to actively mirror our top five trafficked sites at Dreamhost so if Mosso goes down we’ll be able to switch out our nameservers and can be live in production at Dreamhost in an hour.
What do you think? Have a better solution?
Like I said – this is a very rough draft of our cloud disaster recovery plan covering the most mission critical facets of our business. Did we miss something? Have better ideas? Please let us know in the comments!
If you enjoyed this post, make sure you subscribe to my RSS feed!Related posts:
- Budgeting 2009 Web Strategy – Hosting Last year this time we posted about setting up a...
- What Do You Do When Google Is Down? When you’ve come to rely on a service like Skynet...
- Rackspace Announces Cloud Backups Rackspace RackApps announces Cloud Backup and Storage solution for small...
- Windows Azure – A Cloud OS Today, Microsoft jumped on the “cloud computing” bandwagon with its...
- Why Your Business Needs Amazon S3 At 48Web, we use Amazon Simple Storage Service (S3) for...
subscribe to the 
subscribe to the 
follow Andy on 
{ 2 comments… read them below or add one }
Have you thought about using dnsmadeeasy – to auto monitor the mosso webserver and hot-failover to dreamhost if it doesn't respond – within minutes vs waiting for a ns change to propogate and ttl to expire?
That is a FANTASTIC suggestion for mirroring hosting providers