Amazon's east coast cloud crashes, brings down customers

Databases, apps and OSes can't talk; raft of site failures could mar cloud reputation

Enterprise-scale cloud computing showed its dark side overnight, as Amazon's Elastic Compute Cloud developed holes that let several really big customers drop through.

Foursquare, Quora, Reddit, Moby and Hootsuite are among the commercial customers whose web sites failed after Amazon's Infrastructure as a Service began having problems making connections between databases and applications, applications and storage systems and between APIs and everything else.

The reports, the first of which is posted on Amazon's Service Health Dashboard, began at 1:41 a.m. Pacific time cited "latency and error rates with EBS volumes and EC2 instances" across the Eastern U.S.

The site was still experiencing problems for both commercial services and private companies as of 10:40 Eastern time in the U.S. This is the first time EC2 -- the leading example of a stable, flexible-capacity public-cloud service -- has experienced outages serious enough to bring down multiple customers, especially those whose business is as public as high-traffic web sites.

The Wall Street Journal's All Things Digital reports these other outages:

What’s wrong? The new clean desk test
Join the discussion
Be the first to comment on this article. Our Commenting Policies