Which S3 best practice would optimize performance with thousands of PUT request each second to a single bucket?

Amazon S3 has the following structure: S3://BUCKET/FOLDERNAME/FILENAME.zip
Which S3 best practice would optimize performance with thousands of PUT request each second to a single bucket?
A. Prefix folder names with user id; for example, s3://BUCKET/2013-FOLDERNAME/FILENAME.zip
B. Prefix file names with timestamps; for example, s3://BUCKET/FOLDERNAME/2013-26-05-15-00-00-FILENAME.zip
C. Prefix file names with random hex hashes; for example, s3://BUCKET/FOLDERNAME/23a6-FILENAME.zip
D. Prefix folder names with random hex hashes; for example, s3://BUCKET/23a6-FOLDERNAME/FILENAME.zip