Technically, expedited retrieval for files is not guaranteed within 1-5 minutes for files larger than 250 MB+. See https://docs.aws.amazon.com/AmazonS3/latest/userguide/restoring-objects-retrieval-options.html.

S3 Standard-IA is for data that is accessed less frequently, but requires rapid access when needed. S3 Standard-IA offers the high durability, high throughput, and low latency of S3 Standard, with a low per GB storage price and per GB retrieval chargehttps://aws.amazon.com/s3/storage-classes/

"Expedited Retrieval, you can retrieve small amounts of data (up to 250 MB per request) within 1-5 minutes." Cannot C

(A) is eliminated. Demand metric is popularly based & cannot be configured with Lifecycle policies. Ex: an old movie can have resurgent demand 20 years after it's sequel is released. (C) is eliminated. Expedited retrieval is for all but the largest archived objects (250 MB+). (D) is eliminated. Bulk retrieval takes hours. (B) is more expensive than S3 Glacier Flexible Retrieval but it's the only one that works.

AS the pattern is uncertain- Customer could not, in advance, segregate data, the pattern will be determined on the fly - and with regard of the following S3 feature: S3 Intelligent-Tiering is an additional storage class that provides flexibility for data with unknown or changing access patterns. It automates the movement of your objects between storage classes to optimize cost. C will be the most cost effective for this use case.

Expedited 1-5min and for new files intelligent tier is a good option

All old files should be in--> Glacier Flexible Retrieval takes (1-5 minutes) to retrieve the file. New files should not stay in Standard Storage class forever --> Intelligent-Tiering

I don't think C is an option, S3 Glacier Flexible takes hour to retrieve the data. Option A is actually valid, but the way the option A describe it does not consider "demand patterns based on time" So it should be B

There is something I like about option A. It's the only one that deals with what happens with a movie that goes from "new" to "old". With other options, new movies will be new forever.

Option B makes the most sense. Why not option C: 1. This is not an archival use case, the company runs a video streaming service, so objects are still accessed regularly. Accelerated Retrieval is designed for "occasional urgent requests for a subset of archives". 2. The 5 minute timeframe does not apply to items of 250+ MB. 3. Even if the timeframe were valid, it's not guaranteed ("typically") 4. Expedited retrieval is expensive if used frequently ($10.00 per 1,000 requests) - depending on access patterns, this may more than offset the savings in storage costs. https://docs.aws.amazon.com/AmazonS3/latest/userguide/restoring-objects-retrieval-options.html

Option C

Expedited Retrievals (1-5 minutes) Intelligent-Tiering cost

Expedited Retrievals (1-5 minutes) - Intelligent-Tiering cost

I go with C

C -> Expedited Retrievals (1-5 minutes) - Intelligent-Tiering cost (cost effective) D -> Bulk retrievals (5-12 hours) A -> does not consider demand patterns B -> It's ok, but "C" is more good fit to access patterns

option A is most correct option B..for moving files to standard IA , it needs to stay in S3 standard for minimum 30 days. option C..expedited retrieval does not necessarily guarantee big size file retrieval in <=5 minutes. option D... is also wrong as it would take time in hours. sam

Expedited retrievals is typically made available within 1–5 minutes. Each unit of capacity provides that at least three Expedited retrievals can be performed every 5 minutes and provides up to 150 megabytes per second (MBps) of retrieval throughput. There are some limitations but the bottom line is 5 minutes and I believe this leads us to Expedited retrievals. https://docs.aws.amazon.com/amazonglacier/latest/dev/downloading-an-archive-two-steps.html#api-downloading-an-archive-two-steps-retrieval-expedited-capacity