Is storing data in S3 cheaper than paying for cross-az data transfer?

foo

2/7/24, 4:13 PM

Imagine I have two EC2 instances A and B, in the same region/vpc, but in different AZs.

I want to transfer 1TB from A to B.

Now let's imagine two scenarios:

I send the data from A to B directly.
I upload the data from A to S3, then download the data from S3 to B, and finally delete the data from S3.

Pricing-wise:

I get charged at $.01/GB twice (once when data gets out of A, once getting into B), so 1TB * $.01/GB * 2 = $20
Traffic between EC2 and S3 within a region is free, so I get charged, say for storing 1TB for a day. 1TB/day = 33GB/Month. Taking $0.023 as our per-GB-Mo price, that gives us, 33 * $0.023 = $.759.

Is it actually the case that going through S3 is (much, much) cheaper than going direct (even when storing the data for a whole day rather than an hour) or is my analysis flawed?

0 + 3

amazon-s3

amazon-web-services

transfer

Tim

2/7/24, 6:51 PM

Storing 1TB for a month costs about US$23, so storing it for a day costs (23 / 30 = ) $0.78. The S3 pricing page says you pay $0.098 / GB for all data transferred out of S3, though I suspect it's zero rated for VPCs the pricing page doesn't say that, so you should check to be sure.

foo

2/8/24, 10:56 AM

You're correct, I was off by one order of magnitude for S3, which makes the difference between both approaches even larger. it is my understanding that ec2 <-> s3 traffic within the same region, without going over the internet is indeed free

Tim

2/8/24, 6:52 PM

I think EC2 to S3 traffic is zero rated. It could be the situation where the S3 service has retrieval fees separate from bandwidth - I'm not saying that's the case but it's worth considering. The AWS pricing is fairly complex.

Elon Musk

I sit in a Tesla and translated this thread with Ai:

EN: Is storing data in S3 cheaper than paying for cross-az data transfer?

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.