You are designing a site for a new start up which generates cartoon images for people automatically.
Customers will log on to the site, upload an image which is stored in S3. The application then passes a job to
AWS SQS and a fleet of EC2 instances poll the queue to receive new processing jobs. These EC2 instances
will then turn the picture in to a cartoon and will then need to store the processed job somewhere. Users will
typically download the image once (immediately), and then never download the image again. What is the most
commercially feasible method to store the processed images?
A.
Rather than use S3, store the images inside a BLOB on RDS with Multi-AZ configured for redundancy.
B.
Store the images on S3 RRS, and create a lifecycle policy to delete the image after 24 hours.
C.
Store the images on glacier instead of S3.
D.
Use elastic block storage volumes to store the images.
B is the right answer.
The raw image is stored on S3 and it’s very easy to re-generate so S3 RRS is a good choice. And users only download once so using S3-RRS with life cycle is the best choice.