Will not help in this case
1. It was not a full blackout - it was a brownout. You cannot emulate this by shutting down a datacenter
2. The effects on customers are a "Cloudified" (ughh... language vandalism) version of an outage for a centralized SAN/NAS design. You put all of your eggs into one basket you get the expected result. Is the basket called Amazon, EMC, Pure, Tintry or "FairyDust Storage" does not matter. The results are all the same. You cannot fully counter this by having a redundant storage, you have to be redundant by design - something you will never be if you go for a centralized single storage design (there is just way too many ways to nuke that). The question is different - is it worth it for what you are doing to have a centralized storage - you do a cost/benefit analysis vs distributed and if that shows that you can survive commercially an occasional outage you leave the users to scream and shout.