Durability is defined simply as how safe the data is and whether it would be possible to get the data back. How do we say that S3 One Zone-Infrequent access has 11 9’s of durability, given that it is stored in only one zone, there is a possibility of losing the data?
3 Answers
Hi Sundar,
Welcome to the course. That is a great question! Per Amazon, they’ve designed S3 One Zone-IA for 99.999999999% durability for objects in a single Availability Zone. Because S3 One Zone-IA stores data in a single AWS Availability Zone, data stored in this storage class will be lost in the event of Availability Zone destruction. Given that, S3 One Zone-IA should only be used for storing secondary backup copies or easily re-creatable data.
-Kesha
All S3 storage classes have 11 9’s of durability from the documentation (https://aws.amazon.com/s3/storage-classes) except for the S3 RRS. Durability refers to the possibility that the object will remain "intact" after one year. Durability can be imparted by device failures, fires, theft, meteor strikes, earthquakes, or anything that can damage the drives in a data center.
I believe AWS can guarantee 11 9’s for durability in S3 One-Zone-1A because AWS can vouch for its drive quality, security layers of the data centers. The case is different for availability which refers to accessibility and uptime.
Hi Sundar, so remember that an Availability Zone still contains multiple data centers. Data using S3 One Zone-Infrequent access will still be replicated across data-centers within its one availability zone, and still will be available to achieve the 11 9s of durability through these means. However, in an event that a catastrophe occurs such that multiple data centers within an availability zone are destroyed, there is a possibility of the data being lost. However I don’t think this has ever happened.
This what I was thinking. That 11 9’s durability is guaranteed by the fact that the objects are replicated replicated to multiple data centers within a region. This also explains why the availability is 99.5%.
Would it be accurate then to say that S3 standard, being more geographically distributed, likely has better than 11 9’s durability?
All due respect, this does not really answer the question, and I have the same question. The question is how are we justified in saying S3 One Zone has a durability of 11 9s given that our data is not distributed across AZs? To say "the’ve designed [it] for [11 9s] durability for objects in a single [AZ]" does not really answer this. How is it designed to accomplish this level of durability? It is intuitive to say we have a high level of durability given we’ve distributed the data across AZs, but under the assumption that we have not distributed the data across AZs, how are we sure that we have that level of durability? I hope that clarifies the question. I don’t really even know where this measure comes from, but I’ll save that for another discussion.
Greetins, thank you for valuable comment. I also do share same question and cant’t understand how Kesha post answers it. Even more I would like to ask how AWS gets 11 9’s% where are data showing this? Without any info it all looks like marketing "fake numbers".
Actually, it does answer the question. The wording described for One Zone -IA is "Designed for durability of 99.999999999% of objects in a single Availability Zone" with a disclaimer "Because S3 One Zone-IA stores data in a single AWS Availability Zone, data stored in this storage class will be lost in the event of Availability Zone destruction." So the 11 9’s of durability is for a single availability zone