Event
Amazon startet S3
In March 2006 Amazon launched S3 as a storage service with SOAP and REST interfaces. Amazon’s contemporary press release and Jeff Barr’s blog post identify 14 March as the day it became available. A different AWS announcement page bears the date 13 March; the reason for this discrepancy remains unresolved. Press release and comparison Blog · Announcement
At launch Jeff Barr described data blocks up to 5 GB with user-defined keys, metadata and access-control lists. His 99.99 percent availability figure was a design target, not a measured result. Contemporary vendor account
AWS cited 15 US cents per gigabyte per month for storage and an additional 20 US cents per gigabyte transferred. The EC2 service offered later that year instead provided virtual computing capacity. Source
In February 2007 Simson Garfinkel described his own S3 experiments for computer forensics. Over the connections he tested, he measured an average of around 1–2 MB per second; some write requests required retries. He considered the service useful for his research but raised concerns about security and long-term availability. This is an experience report under particular conditions, not a general failure-rate statistic. USENIX report
Sources
Quellenhinweis: Absatz zum Start von Amazon S3.
Offen
Datumsabweichung zwischen AWS-Ankündigung und Blog weiter prüfen. / Investigate the date discrepancy between AWS announcement and blog.