SAA-C03 Practice Question 68: S3 Multipart Large Uploads
AWS Certified Solutions Architect – Associate (SAA-C03) · Test 7 · question 8 of 10
An editing studio uploads 40 GB video masters from its suites to S3 with a tool that issues a single PutObject call per file. The link is fast but drops briefly every few hours, and when that happens the upload restarts from zero and the operators lose most of a day's transfer. They also observe that a single upload never uses more than a fraction of the available bandwidth.
What should be changed in the upload tool?
Show answer and explanation
Answer:
Multipart upload breaks an object into parts that are uploaded independently and in any order, and S3 assembles them into a single object at the end. That fixes both symptoms: parts sent in parallel use the available bandwidth instead of being limited by one stream, and if a part fails it is retransmitted on its own without affecting the others, so a dropped link costs one part rather than 40 GB. AWS recommends multipart upload for objects of 100 MB or more, and it is not optional at this size; a single PutObject is limited to 5 GB.
Option A improves the path the bytes take by routing them over the edge network, but it changes nothing about failure behaviour: a single PutObject that is interrupted still starts again from the beginning, and it still cannot carry a 40 GB object at all.
Option C reinvents multipart upload badly. Each fragment becomes a distinct S3 object, so every consumer needs the reassembly logic and the studio loses the single master object, whereas multipart gives the same parallelism and produces one object.
Option D misunderstands versioning. A version is created when an upload completes; an interrupted PutObject creates nothing to recover. Versioning protects against overwrite and deletion, not against a failed transfer.
Reference: Uploading and copying objects using multipart upload in Amazon S3