I can't see wifi happening, the power budget for 802.11 is out of scale with the budget for BLE. For BLE you can get away with a fairly simple radio and do everything in software. For wifi you need to implement far more of the protocol in hardware (pretty much the entire MAC). I'd imagine they could add hardware support for 802.15.4 (zigbee) without that much of a silicon area hit, it would just require a software investment for the protocol stack.
I'd imagine that the primary complaints they get from customers is 'lower power/more computes/more flash/more sram'. The switch to 55nm will give them a nice boost in the power budget, clock frequency and transistor budget. But they still have the problem that the BLE stack is somewhat hard realtime. So you can't use the nrf51 for timing critical things. They could move more of the protocol stack into hardware or they could just spilt the work. Have a dedicated core for BLE and a dedicated core for the user application. If it was me I wouldn't just throw down a pair of M0+s cores, but would go for the big/little solution. A small/slow/low voltage M0+ for the BLE stack and a larger/faster M3 for the user application.
I can't see wifi happening, the power budget for 802.11 is out of scale with the budget for BLE. For BLE you can get away with a fairly simple radio and do everything in software. For wifi you need to implement far more of the protocol in hardware (pretty much the entire MAC). I'd imagine they could add hardware support for 802.15.4 (zigbee) without that much of a silicon area hit, it would just require a software investment for the protocol stack.
I'd imagine that the primary complaints they get from customers is 'lower power/more computes/more flash/more sram'. The switch to 55nm will give them a nice boost in the power budget, clock frequency and transistor budget. But they still have the problem that the BLE stack is somewhat hard realtime. So you can't use the nrf51 for timing critical things. They could move more of the protocol stack into hardware or they could just spilt the work. Have a dedicated core for BLE and a dedicated core for the user application. If it was me I wouldn't just throw down a pair of M0+s cores, but would go for the big/little solution. A small/slow/low voltage M0+ for the BLE stack and a larger/faster M3 for the user application.
