Of course I have to comment on the AWS outage! I think Hybrid IT is the way to go – you need both on-prem and off-prem to support your apps and data and have load balancing and other jazz in case something like this happens. Data protection, backup, backup for apps even multiple data centers – you've gotta do it!
So in the spirit of fun, here's all I have to say to those on AWS only:
Clearly you know nothing about Cloud or datacentre infrastructure.
The AWS outage wluld be easily mitigated if you had geo-redundancy in another AWS datacentre. For instance the AWS European DCs were absolutely fine and not impacted. If you had this as a DR site with full automated failover there would be no issue.
Hybrid cloud is great if you have existing investments in your own datacentre where you have, say, 2 years to run on your Cisco flex pod lease agreement and are looking for a phased transition to public cloud, but it doesn’t remove the requirement for sensible, industry best practice approaches to system architecture.
At the end of the day if a single datacentre is what you choose, then you have a single point of failure, period.
Yeti
Hi Yeti! First, thanks for coming to my blog and commenting. Appreciate it! Next, this post was meant for giggles – nothing more. And I think we all know it was a matter of time before something like this happened. A lot of apps were down today – coincidence I think not. So I think they didn’t do the geo-redundancy. Further, I think it’s pretty clear that this is a UX and Agile blog. So I posted it as something for fun with my thoughts.
Just to add, Hybrid Cloud can be great if you have some data that you may want to keep inside your firewall for added security. But just a side talking point…