Discussion Forums
Advanced search options
S3 data reliability
Posted by: Mr. M. Boswell RealName(TM)
Posted on: Mar 26, 2008 10:37 AM
  Click to reply to this thread Reply
I am just starting to use S3, and am curious as to the level of guarantee for data integrity.

I am looking to know how reliable I can assume S3 to be, if I upload (and verify) 5gb of data today, what sort of expectations can I have that it will a) still be there b) still be consistent in a year’s time.

I do understand the principles behind the distributed model of S3, I am curious as to whether there is any sort of service level for this – I have looked through the agreement and can’t see anything

Has anyone had any experience of data loss – that was not due to user error.
Permlink Replies: 2 | Pages: 1 - Last Post: Mar 26, 2008 12:51 PM by: M. Garnaat
Replies
Re: S3 data reliability
Posted by: Allen
Posted on: Mar 26, 2008 11:22 AM
in response to: Mr. M. Boswell in response to: Mr. M. Boswell
  Click to reply to this thread Reply
S3 is nearly as reliable as an online system can be.  There is no known case of data loss that was not caused by user or tool error.
Re: S3 data reliability
Posted by: M. Garnaat
Posted on: Mar 26, 2008 12:51 PM
in response to: Mr. M. Boswell in response to: Mr. M. Boswell
  Click to reply to this thread Reply
Usually, when people ask for SLA's or guarantees what they really are asking for is some assurance that their data will not be lost.  The best assurance of future reliability is past performance and in this regard I think Allen is right.  As far as I know, no one has lost data in S3 because of a failure in the system.  That's over a three year period with billions and billions of objects.  I'll take that over a contract any day of the week.

Mitch