Create better IDs for S3 Records

Question

Create better IDs for S3 Records

bjinwright opened this issue 8 years ago · 3 comments

Check out the following article and YouTube video to learn why this is important for performance.
http://docs.aws.amazon.com/AmazonS3/latest/dev/request-rate-perf-considerations.html#introduce-randomness-hash and https://youtu.be/x3wxVCvqpa8?t=47m29s

Answer 1 · 2016-09-20T13:31:47.000Z

So it looks like the hash is really important for hardware distribution. At the end of the Q&A the guy says you need a secondary index if you want to use hash keys (which makes sense). Rather than use a DynamoDB as the secondary index, could a single document, or small set of documents, in S3 (maybe behind cloudfront) serve the same purpose with less cost? Further it keeps it pure S3 which I like for simplicity sake.

Answer 2 · 2016-09-20T15:55:18.000Z

Check out my answer here. #10 (comment)

Answer 3 · 2016-09-20T16:00:33.000Z

This is fixed in 493741a