Skip to content

Problem with very low data transfer only to AWS SRV (all locations) - the problem concerns the source IP only my pool of external addresses

0

Subject: Degraded TCP/443 upload performance from specific IP prefix to AWS

Description:

We are observing consistently very low upload throughput over TCP/443 from our public IP prefix to multiple AWS endpoints. Download performance is normal; only uploads are affected.

Key facts:

Issue affects only our public IP range (single IPv4 prefix).

When testing from other public IP ranges, upload performance to the same AWS endpoints is normal (20–30x higher).

The issue occurs across multiple AWS IPs and services, not a single destination.

The problem has been present since August 2025; before that, performance was normal.

Network tests:

Tests were performed on two independent ISP links, in different locations.

The issue persists even when traffic is sent directly from the operator edge, with all UTM/firewall devices removed from the path.

ISPs report no traffic shaping or policies applied specifically to our prefix.

UDP tests to AWS show normal upload throughput, while TCP/443 uploads are capped at approximately 2–3 Mbps.

This behaviour suggests possible:

throttling, filtering, or protection policies applied to our IP prefix,

reputation-based rate limiting, or

routing / edge handling issues for TCP traffic originating from our range.

Request:

We kindly ask AWS to:

Verify whether our public IP prefix is subject to any throttling, filtering, or protection mechanisms (e.g. DDoS mitigation, reputation-based rate limits, WAF/Shield policies).

Check if our prefix appears in any internal risk/abuse lists.

Review routing and edge handling for TCP/443 traffic originating from our prefix toward AWS endpoints.

We can provide:

exact source prefix (privately, if required),

example destination AWS IPs,

traceroute/MTR results,

comparative measurements from other IP ranges.

2 Answers
0
Accepted Answer

You should absolutely open a support case. re:Post is a community forum and folks here cannot help look at specific network issues from the AWS side.

AWS
EXPERT

answered a year ago

  • Unfortunately, I don't have active business services in AWS to create a technical ticket, so I decided to ask here :)

0

Based on your detailed description, it appears you're experiencing a specific network performance issue affecting TCP/443 uploads from your IP prefix to AWS services. This asymmetric behavior (normal downloads but severely limited uploads) is unusual and suggests several possible causes.

Since the issue:

  • Only affects your specific IP prefix
  • Persists across different ISPs and locations
  • Occurs even with network equipment removed
  • Started at a specific point in time (August 2025)
  • Only impacts TCP/443 uploads while UDP performs normally
  • Shows normal performance when testing from other IP ranges

This strongly suggests some form of traffic management or filtering specifically targeting your IP range. Here are the most likely possibilities:

  1. Reputation-based filtering: AWS may have flagged your IP prefix in their threat intelligence systems. This could happen if the IP range was previously associated with abusive behavior, even if it wasn't from your organization.

  2. DDoS protection mechanisms: AWS Shield or similar protection services might be applying throttling to your prefix based on historical traffic patterns.

  3. TCP-specific network path issues: The fact that UDP works normally while TCP/443 is affected points to possible TCP window size issues or TCP-specific filtering along the network path.

  4. Regional routing problems: There could be suboptimal routing between your network and AWS edge locations, causing TCP performance degradation over high-latency paths.

To resolve this issue, I recommend:

  1. Contact AWS Support directly with your specific IP prefix information. They can check internal systems for any reputation flags or filtering policies.

  2. Request AWS to review logs for any automated protection systems that might be throttling traffic from your prefix.

  3. If possible, try testing with different TCP ports to see if the issue is specific to port 443 or affects all TCP traffic.

  4. Consider implementing AWS S3 Transfer Acceleration if you're transferring data to S3, as this uses different network paths through AWS edge locations.

  5. Try adjusting TCP window sizes and other TCP optimization parameters if you have control over the sending systems.

This appears to be a specific case that will require direct investigation by AWS network engineers with access to their internal systems.
Sources
The outbound speed from us-east-1 to Hong Kong is super slow. Serving website and download using Winscp is slow. The upload is normal or close to normal though. | AWS re:Post
S3 Transfer Rates Capped at 2.8mb/s ??? How Can I Speed This Up? | AWS re:Post

answered a year ago

AWS
EXPERT

reviewed a year ago

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.