On our iPhone 15 Pro running iOS 27.2 (24B119), a single physical BLE peripheral consistently appears as two separate scan results with different UUIDs. The issue reproduces consistently when the peripheral uses connectable BLE extended advertising with a 1M primary PHY and 2M secondary PHY. Switching the same peripheral to legacy advertising eliminates the duplicate scan results. The duplication is visible in both our application and nRF Connect, and we did not observe it on iOS 26.
The duplicate results are visible in both our application and nRF Connect, so this is not specific to our application code or BLE library.
The peripherals use a Silicon Labs BLE chipset and connectable BLE extended advertising with a 1M primary PHY and 2M secondary PHY.
The interesting detail is in the captured bluetoothd logs: the same Bluetooth address, advertised local name, and manufacturer data appear under two UUIDs, but one identity is labeled Random and the other Public. Both are marked connectable.
For example, one peripheral appears as:
Local name: WD-883200239
Bluetooth address: 7C:C6:B6:58:F4:76
Random:
F015AD07-F340-40DB-CEB5-5D602386B408
Public:
6D0F1C74-14B9-06B3-F64C-29E33CB8A065
On November 6, the logs report the Random identity at 08:07:07.496, followed by the Public identity at 08:07:08.505, with matching address bytes, local name, and manufacturer data. A second physical peripheral shows the same pattern.
To clarify, these are different UUIDs for the same physical peripheral, not simply repeated discovery callbacks for one UUID.
This creates ambiguity when selecting a peripheral for connection. The Public/Random distinction is visible in system diagnostics but is not exposed through supported Core Bluetooth APIs. Both identities are marked connectable, so that flag does not help select the usable identity. We have not yet established that one address type consistently succeeds while the other fails.
I have filed FB25543615 with the sysdiagnose and Bluetooth HCI PacketLogger captures.
I’m posting separately from the extended-advertising discovery issue identified as r.188123531, which Apple clarified affects iPhone 17 and newer. This reproduction is on an iPhone 15 Pro and concerns duplicate identities, rather than missing discovery. I’m not assuming the two issues share a cause.
Has anyone else observed this after moving from iOS 26 to iOS 27? In particular:
- What could cause the same address to be reported as both Public and Random, producing separate Core Bluetooth UUIDs?
- Is there a supported way for an application to distinguish the usable identity before attempting a connection?
- What additional controller traces or over-the-air evidence would help determine where the address-type inconsistency originates?