Exams / SOA-C03 / Study guide / S3 performance and data transfer

S3 performance and data transfer: SOA-C03 study notes and practice questions

What to know about S3 performance and data transfer for the SOA-C03 exam, from the Monitoring, Logging, Analysis, Remediation, and Performance Optimization domain (22% of the exam guide), then 1 free question to test it. Part of the SOA-C03 study guide.

Last verified October 2026 against current AWS documentation ·How we write our questions

What to know for the exam

These questions ask how to make S3 uploads, downloads or request rates faster, or how to move a lot of data.

Uploads and downloads

Multipart upload splits a file into parts that upload in parallel and retry individually. Required above 5 GB (single PUT limit); recommended from about 100 MB. The AWS CLI uses it automatically for large files.
Add a lifecycle rule to abort incomplete multipart uploads, or orphaned parts keep billing.
S3 Transfer Acceleration sends data through the nearest edge location and the AWS network; it helps distant clients. Bucket names must be DNS-compliant without dots, and clients use the accelerate endpoint.
Byte-range fetches download parts of an object in parallel, or only the part you need.

Request rates

S3 supports at least 3,500 write and 5,500 read requests per second per prefix, with no limit on prefixes. If an application gets 503 Slow Down errors, spread keys across more prefixes and retry with backoff.

Moving data

Tool	Use
AWS DataSync	Online transfers from NFS, SMB, HDFS, other clouds or between AWS storage, with scheduling, verification and bandwidth limits
AWS Snowball Edge	Offline transfer of terabytes to petabytes when the network is too slow
S3 Batch Operations	Run an action (copy, restore, tag, invoke Lambda) on millions of existing objects
S3 Replication	Copy new objects to another bucket or Region automatically

Watching performance

S3 request metrics in CloudWatch (per bucket, opt-in) show latency, errors and request counts. S3 Storage Lens gives organization-wide usage and activity.

Common traps

Transfer Acceleration for repeated downloads of the same files. Use CloudFront.
Retrying a failed 50 GB upload in one PUT. It can’t work; use multipart.
DataSync vs Snowball. Enough bandwidth points to DataSync; weeks of transfer time points to Snowball.

Try the 1 practice question

Practice S3 performance and data transfer Take the free practice exam

Question 1

Medium · choose 2 ·Link to this question

An esports league records match footage at venues on several continents and uploads each file, between 8 GB and 20 GB, to one S3 bucket in us-east-1.

Uploads from distant venues are slow, and a single network interruption forces the whole file to be sent again.

The league wants faster, more resilient uploads without changing the bucket's Region.

Which TWO actions should a CloudOps engineer recommend? (Choose two.)

Choose two answers.

A. Upload each file with multipart upload so that parts transfer in parallel and only failed parts are sent again.
B. Turn on S3 Transfer Acceleration for the bucket and upload through the accelerate endpoint from each venue.
C. Turn on S3 Versioning so that an interrupted upload resumes from the last version that was stored in the bucket.
D. Upload the footage to an S3 directory bucket that uses S3 Express One Zone, which is built for lower-latency writes.
E. Spread the files across many key prefixes so that S3 gives each prefix its own request rate limit.

Show the answer and explanation

Answers: A, B

A. Right: Correct. Multipart upload splits a file into parts that upload in parallel and can be retried one at a time. A single PUT is limited to 5 GB, so files of this size need multipart upload anyway. Learn more: Uploading and copying objects using multipart upload (opens in a new tab)
B. Right: Correct. Transfer Acceleration routes uploads into the nearest CloudFront edge location and then across the AWS network to the bucket. It helps most when clients are far from the bucket's Region. Learn more: Configuring fast, secure file transfers using Amazon S3 Transfer Acceleration (opens in a new tab)
C. Wrong: Versioning keeps earlier versions of objects that are overwritten or deleted. An interrupted upload does not create a partial version, so nothing can be resumed this way. Learn more: Retaining multiple versions of objects with S3 Versioning (opens in a new tab)
D. Wrong: S3 Express One Zone stores data in a single Availability Zone and gives low latency to compute in that same zone. It does not speed up long-distance uploads from venues and changes how the data is stored. Learn more: Understanding and managing Amazon S3 storage classes (opens in a new tab)
E. Wrong: Prefixes raise the number of requests per second a bucket can handle. These uploads are a few very large files, so request rate is not the bottleneck. Learn more: Best practices for Amazon S3 performance (opens in a new tab)

Explanation

Large files over long distances need two fixes. Multipart upload sends parts in parallel and retries only what fails, and it is required above 5 GB. Transfer Acceleration shortens the slow public internet part of the path by entering the AWS network at a nearby edge location. Prefixes, storage classes and versioning do not address either problem.

Study

Want more on S3 performance and data transfer? The 5 full-length practice sets have 5 more S3 performance and data transfer questions, 325 in all, none of them in the free exam. One payment of $14.99, no subscription. See the practice sets

Study S3 performance and data transfer

Best practices for Amazon S3 performance