Amazon S3 – 2 Trillion Objects, 1.1 Million Requests/Second • Hacker News Board

Amazon S3 – 2 Trillion Objects, 1.1 Million Requests/Second

#1

Post by jeffbarr » Thu, Apr 18, 2013, 2:00 PM UTC

Amazon S3 – 2 Trillion Objects, 1.1 Million Requests/Second
aws.typepad.com

Re: Amazon S3 – 2 Trillion Objects, 1.1 Million Requests/Second

#2

Post by MartinMond » Thu, Apr 18, 2013, 2:10 PM UTC

So how many objects should have been corrupted/lost according to their SLA and how many actually did get corrupted?

Re: Amazon S3 – 2 Trillion Objects, 1.1 Million Requests/Second

#3

Post by Smrchy » Thu, Apr 18, 2013, 2:13 PM UTC

It would be really interesting to know the average size of an object and visualize the amount of harddisks it takes to store all this data.

Re: Amazon S3 – 2 Trillion Objects, 1.1 Million Requests/Second

#4

Post by HeyImAlex » Thu, Apr 18, 2013, 2:21 PM UTC

MartinMond wrote: post #2

So how many objects should have been corrupted/lost according to their SLA and how many actually did get corrupted?

>how many objects should have been corrupted/lost according to their SLA

Anywhere from 20 to 200M per year depending on how many people use RRS...

Re: Amazon S3 – 2 Trillion Objects, 1.1 Million Requests/Second

#5

Post by cinquemb » Thu, Apr 18, 2013, 2:25 PM UTC

waiting for the comparison to the total outstanding us derivative exposure :P

Re: Amazon S3 – 2 Trillion Objects, 1.1 Million Requests/Second

#6

Post by jeffbarr » Thu, Apr 18, 2013, 2:28 PM UTC

cinquemb wrote: post #5

waiting for the comparison to the total outstanding us derivative exposure :P

Dollars and objects don't compare very well.

Re: Amazon S3 – 2 Trillion Objects, 1.1 Million Requests/Second

#7

Post by jeffbarr » Thu, Apr 18, 2013, 2:30 PM UTC

HeyImAlex wrote: post #4

MartinMond wrote: post #2

So how many objects should have been corrupted/lost according to their SLA and how many actually did get corrupted?

>how many objects should have been corrupted/lost according to their SLA Anywhere from 20 to 200M per year depending on how many people use RRS...

Note that RRS is "Reduced Redundancy Storage" - an S3 option that offers 99.99% durability, as opposed to the 99.999999999% durability offered by Standard S3.

Re: Amazon S3 – 2 Trillion Objects, 1.1 Million Requests/Second

#8

Post by TallboyOne » Thu, Apr 18, 2013, 2:34 PM UTC

What is an 'object' in this sense?

Re: Amazon S3 – 2 Trillion Objects, 1.1 Million Requests/Second

#9

Post by chuckmans3 » Thu, Apr 18, 2013, 2:34 PM UTC

Considering we don't hear about problems that often, this is quite an impressive feat of engineering. You really don't think about it until we get a hiccup and half the Internet goes down.

Re: Amazon S3 – 2 Trillion Objects, 1.1 Million Requests/Second

#10

Post by ihsw » Thu, Apr 18, 2013, 2:36 PM UTC

Smrchy wrote: post #3

It would be really interesting to know the average size of an object and visualize the amount of harddisks it takes to store all this data.

To put this into perspective it's 2.7B new objects per day (assuming 1 Trillion objects averaged over 365 days).

Assuming each object is 100KB (generous estimate, after compression) that would be 270GB per day -- or assuming ten levels of redundancy and striped across three RAID storage devices (per level of redundancy) then 8.1TB per day.

I'm not familiar with their hard disk procurement policies but it wouldn't be difficult to assume they've been purchasing 1TB drives, so 10 new disk drives per day just for keeping ahead of growth. Furthermore let's assume their disk drive failure churn rate is 10% per day so another 1 new disk drive for parts replacement (so 11 disk drives per day).

These are really loose numbers not based on any actual data (or any personal experience at all) but just napkin math, so take it all with a grain of salt.