AWS latency test
Read your latency results
Your browser measures HTTPS request-response times to Amazon S3 in selected regions. Lower median latency is better.
Early values show the first responses. They are provisional and do not enter the stable ranking. A median result needs at least three successful samples after warm-up.
Quick comparisons run 10 rounds. Live sampling continues for more rounds and can be paused at any time. Queued regions are waiting for a request slot.
Unavailable means this connection could not measure the region. Check your connection or retest; it does not mean that the cloud region is down.
Use the results to shortlist regions. Application behavior, cost, service availability, data residency and resiliency also affect your choice.
For TCP connection measurements, read the PsPing guide and use your AWS endpoint as the target.
Find the lowest-latency AWS region for your connection.
Select regions
Compare all 34 regions in 10 rounds.
0 regions Choose regions to start Change regions
Latency ranking
No regions selected.
No latency data yet
Choose AWS regions to start measuring latency from your connection.
Frequently asked questions
What is latency and what constitutes good latency?
Latency, often referred to as ping, is the duration for data to travel from the source to the destination and back. The results of this test represent the median round trip time (RTT) latency to AWS storage endpoints, with lower RTT indicating superior performance.
What constitutes good latency? Good latency varies by application type:
- Real-time applications (e.g., gaming, video conferencing): latency below 50 ms is ideal.
- Interactive applications (e.g., web browsing, online trading): latency between 50 ms and 100 ms is usually acceptable.
- Non-interactive applications (e.g., file transfers, backups): latency above 100 ms may be acceptable.
These guidelines may vary depending on specific application needs and user expectations.
How does AWS Latency Test work?
Your browser sends HTTPS requests to AWS storage files in each region. The median latency is calculated by measuring the time between the request and the response.
If measurements fail, check your connection and retest. Previous successful results stay visible but are excluded from the live ranking until the region responds again.
Does the latency test reflect actual application performance?
Partially. This test is a network-focused indicator, not a full application benchmark.
- Good for comparing relative latency between AWS regions
- Useful for region-selection discussions
- Not a substitute for application-level load or performance testing
Use this test as a signal for initial guidance, not as the sole decision-maker.
Is my speed test result private?
Yes. AWSSpeed does not require authentication and does not collect personal or corporate identity information. Test results are generated entirely in your browser and are only visible to you. Other users cannot view, discover, or search for your results unless you explicitly share them (e.g., via screenshot). Results are not published, stored, or retained on our servers.
Do you support upload, download tests, packet loss, and jitter?
Currently not supported.
Why might my AWS Latency Test result differ from my PsPing result?
Browser-based latency tests include additional overhead that PsPing does not measure. This test uses HTTPS requests which include:
- DNS resolution time
- TLS/SSL handshake negotiation
- HTTP request/response header processing
- Browser JavaScript execution overhead
PsPing measures raw TCP or ICMP echo requests with minimal overhead. Expect browser results to be higher than PsPing. Both are valid — this test reflects real-world browser application latency, while PsPing shows pure network latency.