A company runs a photo hosting service in the us-east-1 Region. Users from multiple countries upload and view photos. Some photos are heavily viewed for months; others are viewed for less than a week. Each photo upload can be up to 20 MB. The service uses photo metadata to determine which photos to display to each user. Which solution provides appropriate user access MOST cost-effectively?

The correct solution leverages S3 Intelligent-Tiering for cost-effective storage of photos with varying access patterns. Intelligent-Tiering automatically moves objects between access tiers based on usage, optimizing costs without performance impact, which is ideal for photos with unpredictable access. Storing metadata in DynamoDB is efficient for fast, low-latency access to photo details and their S3 locations. Storing photos in DynamoDB is incorrect because DynamoDB is a NoSQL database designed for small, structured data, not large binary objects like photos. S3 Standard with a lifecycle policy to S3 Standard-IA is less optimal than Intelligent-Tiering because it requires manual policy configuration and doesn't adapt to changing access patterns as effectively. S3 Glacier is designed for archival data with infrequent access and higher retrieval times, making it unsuitable for frequently viewed photos. Using OpenSearch Service for metadata is an over-engineered and more expensive solution compared to DynamoDB for simple metadata storage and retrieval.