As part of their new application experience, Dress4Wm allows customers to upload images of themselves. The customer has exclusive control over who may view these images. Customers should be able to upload images with minimal latency and also be shown their images quickly on the main application page when they log in. Which configuration should Dress4Win use?

Storing image files in a Google Cloud Storage bucket is ideal for this use case because Cloud Storage offers high availability, scalability, and low latency for uploads and retrievals, which are crucial for a good customer experience. Using Google Cloud Datastore (now Firestore in Datastore mode) for metadata is appropriate because it's a NoSQL document database that scales automatically, handles semi-structured data well, and provides fast lookups for mapping customer IDs to their image files. This combination ensures both efficient storage of large binary objects (images) and quick, scalable access to their associated metadata. The option to add custom metadata to Cloud Storage objects is less efficient for querying than a dedicated database. Distributed file systems are generally more complex to manage and scale than Cloud Storage, and assigning file owner attributes for privacy is not a standard or scalable solution in a cloud environment. While Cloud SQL could store metadata, Datastore is often preferred for its automatic scaling and flexible schema, which suits evolving application needs better than a relational database for this type of metadata.