AUTHOR: http://aws.typepad.com/aws/2011/10/new-whitepaper-use-aws-for-disaster-recovery.html LINK!

Recent AWS Customer Success Stories & Videos

More AWS Customer Success Stories...

« New - AWS Management Console Support for the Simple Email Service | Main | New - AWS Virtual (Software) Multi-Factor Authentication - RFC 6238 Support »

TrackBack

TrackBack URL for this entry:
http://www.typepad.com/services/trackback/6a00d8341c534853ef0153928fa06a970b

Listed below are links to weblogs that reference New Whitepaper: Using AWS for Disaster Recovery:

Comments

Feed You can follow this conversation by subscribing to the comment feed for this post.

T.Pham

Do you have a warranted recovery time objective on your cloud disaster recovery offerings?

Brian Adler

Hi Jinesh,
Thanks very much for the info provided here. After reviewing the white paper, what I don’t see are recommendations on configuring a DR environment in AWS in which the primary environment is also in AWS. The paper discusses the on-premise scenario with the DR environment in the cloud, which is obviously a viable use case, but the majority of customers we talk to about a disaster recovery solution (full disclosure: I am an Architect in the Professional Services group at RightScale) are running in AWS and want to set up a DR environment in another AWS region. Currently there aren’t any easy ways to replicate a database between two AWS regions – if you use a relational database in EC2 (for example, and of those discussed in the link on page 7), you are most likely going to be using EBS volumes for your datastore. The snapshots you take of these EBS volumes are not portable across regions, nor are they discoverable as objects in an S3 bucket so they could be copied to another bucket in a different region. Similar constraints exist with RDS in that EBS volumes are used for storage (so similar issues exist with regard to the snapshots and their portability), and there is currently no way to configure a slave to replicate from an RDS instance, so keeping a database replica current in another region is not really possible if using RDS. We have tips and tricks we have used to enable database replication for both ourselves and our customers, but given some of the AWS region-portability constraints, we find it challenging to provide an optimal solution 100 percent of the time. I think this white paper is a great start, but it would be really useful to see additional details on recommendations and best practices for setting up an AWS region as a DR environment for another AWS region. Thanks again for the post and the white paper, as the information provided is indeed relevant for some use cases.
- Brian Adler, PS Architect, RightScale

Scott McDonald

Hi Brian,

FWIW - it can be done easily with traditional lamp stacks running as persistent instances. MySQL master/master replication can be setup between 2 servers in different AWS regions - combined with simple automated recurring rsync's to keep static web content (from user uploads etc...) in sync between the two AWS regions. Combined with DNS failover techniques to provide HA - it works great.

I'm not a proponent of using multiple AWS regions for HA though. I think it is much better to have HA setup between a cloud in a single AWS region and a cloud from a completely different vendor (i.e. Linode or Rackspace). Vendors have vendor wide issues. Best to have redundant vendors in your HA setups as well. That's what I do with DNSHAT.com and what I recommend for DNSHAT.com failover clients that are looking to automate failover between different cloud systems.

Thanks,
Scott McDonald - Founder - DNSHAT.com

The comments to this entry are closed.

Featured Events

The AWS Report


Brought to You By

Jeff Barr (@jeffbarr):



Jinesh Varia (@jinman):


Email Subscription

Enter your email address:

Delivered by FeedBurner

April 2014

Sun Mon Tue Wed Thu Fri Sat
    1 2 3 4 5
6 7 8 9 10 11 12
13 14 15 16 17 18 19
20 21 22 23 24 25 26
27 28 29 30