Optimize ESP32 OTA Delivery

I keep all firmware binaries for our ESP32 fleet in an S3 bucket housed in us-east-1. Each board receives a presigned HTTPS link and pulls the binary through the ESPIDF OTA routine. In Sri Lanka everything flashes first-try, however units deployed in Sweden stall; the transfer takes so long that the built-in timeout fires and the update aborts.

To narrow the gap I placed a CloudFront distribution in front of the bucket and saw a noticeable speed-up during local testing, yet the European devices still can’t finish a >5 MB image over their Wi-Fi connection. So the core issue—excessive latency and/or throughput limits between the edge location and the S3 origin—remains.

I need a clean AWS-side solution that guarantees every ESP32 can fetch the full image within the default OTA time-out, without me having to rebuild the firmware logic. Options I’m considering (and happy to validate or discard with your guidance) include:

• Multi-region S3 replication with origin failover or latency routing
• CloudFront Regional Edge Caches or S3 Transfer Acceleration
• Signed cookies or query strings to keep our current authentication flow intact
• Range-request or chunked transfer tuning to match the ESP32’s 4 KB buffer size

This is very urgent for me so I need to get this resolved as soon as possible.