I’m just reading about this now, but apparently Yesterday, 4/22/2011, morning a network event took severely hindered Amazon’s EC2 cloud. According to an article I found on Internetnews.com Amazon is blaming the outage on storage issues caused by a network event.
A report form amazon said:
A networking event early this morning triggered a large amount of re-mirroring of EBS [Extended Block Store] volumes … This re-mirroring created a shortage of capacity … which impacted new EBS volume creation as well as the pace with which we could re-mirror and recover affected EBS volumes
So, essentially, their storage mechanizm DDOS’d itself. I guess that makes sense. I could speculate here. Imagine that their cloud storage has redundant links, and when some of those links fail, a mirror process is triggered to get the data off somewhere that its accessible, or at least safe. Enough of those links go down, and you have a condition that could cause the issue described above.
If you read my blog, you know i’m no fan of “The Cloud”, so you can probably guess that I’m actually sort of enjoying this news. Just goes to show that resilliant as it is, the concept of cloud computing is not the magic solution that everyone thinks it is.