NRF51-OCB-PY

A port of OCB implementation for NRF51 using softdevice. the test.py allows to receive data from uart and decrypt it. Same py implementation can be used in the cloud, go figure.

Note: there is a Python OCB implementation, however, it supports only version 2. This application supports version 3, hence python example uses ctypes to speed-up crypto and reuses same (ok, almost the same) c implementaiton as on node.

This applicaiton is not suitable for production, not code optimization have been done.

Getting

You need to include submodules as well. I desided to fork and clean tinyAES to minimize the code size. We have only AES ECB mode hence, the other mode not needed. The NRF51 provides encryption, hence we need only soft decrypt

git clone --recursive https://github.com/Northshoot/NRF51-OCB-PY.git

Usage

Compile shared C library

gcc -o ocb_encrypt_lib.so -shared -fPIC -lcrypto  ocb_shared_lib.c

Compile and upload nrf application

make && nrfjprog program -s /PATH_TO_SDK/s110_nrf51822_7.1.0/s110_nrf51822_7.1.0_softdevice.hex -c _build/nrf51422_xxac.hex

tests

Security test zero vectors

data size bytes	encrypt (HW-AES) ms	decrypt (SW_AES) ms	ration SW/HW
32	0.969	1.594	1.64
64	1.53	3.656	2.39
128	2.594	9.25	3.57
256	5.219	18.281	3.50
512	10.062	34.906	3.47
1024	19.031	69.406	3.65
2048	36.25	138.562	3.82
4096	70.688	276.969	3.92

needed improvements

Dynamic array length:

right now these values in C library

#define KEYSIZE  16
#define DATASIZE  32
#define TAGSIZE  16
#define CIPHERSIZE  DATASIZE+TAGSIZE

must be same and manualy edited as in python

KEYSIZE = 16
DATASIZE = 32
TAGSIZE = 16
CIPHERSIZE = DATASIZE+TAGSIZE
KEYBYTES = c_uint8*KEYSIZE
DATABYTES = c_uint8*DATASIZE
CIPHERBYTES = c_uint8 * CIPHERSIZE

class EncryptedData(Structure):
    _fields_ = [
        ("datalength", c_uint32),
        ("key", KEYBYTES),
        ("nonce", KEYBYTES),
        ("assoc", DATABYTES),
        ("cipher", CIPHERBYTES),
        ("cleartext", DATABYTES)]

would be nice to have "automagical" allocation

Copyright

the origial OCB is patented and subject to copyright. Please refere it for usage http://web.cs.ucdavis.edu/~rogaway/ocb/license.htm

The extention and usage in this project is under License 1 — License for Open-Source Software Implementations of OCB (Jan 9, 2013) Under this license, you are authorized to make, use, and distribute open-source software implementations of OCB. This license terminates for you if you sue someone over their open-source software implementation of OCB claiming that you have a patent covering their implementation.

Links

http://web.cs.ucdavis.edu/~rogaway/ocb/ https://en.wikipedia.org/wiki/OCB_mode

Thanks

Huge thanks for Henry & Neil for giving ideas in the dark moments!