How to use a computer as an IoT device to connect AWS IoT service?

Although we don't have Amazon IoT Button or other devices to try AWS IoT, please remember that our computers are devices. Therefore make them IoT.

The sample basicPubSub.py in AWS IoT Python SDK can help us do it.

We follow AWS IoT tutorial to create a device certificate. We download the device certificate, its root CA, and its private key in our local path where the sample exists.

We also create a policy, and a device. We attach both resources into the device certificate.

Because the command arguments of the sample are too long, I prefer to write a makefile Test.mk to test it.

E1 = a2uc??????????.iot.us-west-2.amazonaws.com 
R1 = VeriSign-Class\ 3-Public-Primary-Certification-Authority-G5.pem
C1 = 5988??????-certificate.pem.crt
K1 = 5988??????-private.pem.key

dev1:
    python3 basicPubSub.py -e $(E1) -r $(R1) -c $(C1) -k $(K1)

Run the sample.

>make -f Test.mk dev1

It will publish messages and receive subscribed messages in loop.


-Count

Install AWS IoT Python SDK in Mac

I followed the steps in the page to install AWS IoT Python SDK in my MacBook but I occurred problem.

I checked my Python is version 3.

>python3 -V
Python 3.5.2

But the OpenSSL used by the Python is not 1.0.x that is required by the SDK.

>python3
>>>import ssl
>>>ssl.OPENSSL_VERSION
'OpenSSL 0.9.8zh 14 Jan 2016'

I used brew to download the new OpenSSL.

>brew update
>brew install openssl
>brew link —force openssl

The above command occurred warning.



The command cannot built a link in /usr/bin/openssl because the position is occupied by a real openssl file. I wanted to remove the below openssl but I couldn't.

>cd /usr/bin
>sudo rm -rf openssl
rm: openssl: Operation not permitted

I followed the page to remove it.

>csrutil disable
>reboot
>cd /usr/bin
>sudo rm -rf openssl
>csrutil enable
>reboot

At that time, I made sure that the old openssl was killed.

I built OpenSSL links.

>ln -s /usr/local/opt/openssl/lib/libcrypto.1.0.0.dylib /usr/local/lib/
>ln -s /usr/local/opt/openssl/lib/libssl.1.0.0.dylib /usr/local/lib/
>ln -s /usr/local/Cellar/openssl/1.0.2j/bin/openssl /usr/bin/openssl

I checked the Python's OpenSSL again, it was still 0.9.

>python3
>>>import ssl
>>>ssl.OPENSSL_VERSION
'OpenSSL 0.9.8zh 14 Jan 2016'

The python was installed in a PKG way and the OpenSSL 0.9.8 is embedded in the python. Therefore I wanted to remove the python.

I followed the page to uninstall the old Python 3.

1. Goto Finder>Applications>Python 3.0. Right click, select Move to Trash.

2.
>cd /Library/Frameworks/
>sudo rm -rf Python.framework

I used brew to install Python 3 with OpenSSL 1.0.2.

>brew install python3 --with-brewed-openssl

I checked the Python's OpenSSL again.

>python3
>>>import ssl
>>>ssl.OPENSSL_VERSION
'OpenSSL 1.0.2j  26 Sep 2016'

OKAY. I had upgrated Python and OpenSSL.

Then I downloaded the SDK and run a sample in my MacBook. The final steps followed the end of the page.

-Count

Install AWS IoT Python SDK in Windows

AWS IoT is an IoT platform provided by Amazon to connect devices in cloud. We can use the AWS IoT Python SDK to write Python script to access the AWS IoT platform through MQTT to interact with connected devices. The following are steps of installing the SDK.

Check if the current Python is version 3.

>python -V
Python 3.5.2 ...

I prefer to use Python 3. If it is not, we can download the newest Python 3.5.2 from the website and install it.

https://www.python.org/ftp/python/3.5.2/python-3.5.2.exe

Check the OpenSSL version used by the Python.

>>> import ssl
>>> ssl.OPENSSL_VERSION
'OpenSSL 1.0.2h  3 May 2016'

Install the AWS IoT Python SDK from source that contains samples.

>git clone https://github.com/aws/aws-iot-device-sdk-python.git
>cd aws-iot-device-sdk-python
>python setup.py install

Indeed, we can also install the SDK from pip, but it doesn't have samples.

>pip install AWSIoTPythonSDK

Run a sample

>cd samples/basicPubSub
>python basicPubSub.py

An error occurs.



This page has a workaround:

https://github.com/aws/aws-iot-device-sdk-python/issues/23
import os
import sys
import AWSIoTPythonSDK
sys.path.insert(0, os.path.dirname(AWSIoTPythonSDK.__file__))
# Now the import statement should work
from AWSIoTPythonSDK.MQTTLib import AWSIoTMQTTClient

Run the sample again.

>python basicPubSub.py -h



-Count

Why cannot S-box of DES cipher be linear?

We cannot specify any S-box in DES cipher. One of the rules to choice of S-box is that, it cannot be linear. How do we understand it? Below is my explanation, but I'm not sure it is correct.

The Stanford on-line course, Cryptography I, describes that, if S-box is chosen as linear, the DES cipher would be linear.

DES(k, m) = B x |m  |= c    (statement 1)
                |k1 |
                |k2 |
                |.  |
                |.  |
                |.  |
                |k16|

where:
|m| = |c| = 64 (bits)
|ki| = 48 (bits)
B is a 64x832 matrix.

The course describes that, "You just need 832 input output pairs, and you'll be able recover the entire secret key." How to explain the description? We can use a simple example:

Lets
B is 2 x 4 matrix.
|m| = |c| = |k| = 2 (bits)

We transform the statement 1 as below.

|b1 b2 b3 b4| x |m1| = |c1|
|b5 b6 b7 b8|   |m2|   |c2|
                |k1|
                |k2|

We give the below statement.

b1m1 xor b2m2 xor b3k1 xor b4k2 = c1    (statement 2)
b5m1 xor b6m2 xor b7k1 xor b8k2 = c2    (statement 3) 

We assume that, we don't know any b and any k, but we know the input m and output c.

For statement 2, the number of unknown variables is 4. They are,
b1, b2, b3k1, b4k2.

For statement 3, the number of unknown variables is 4. They are,
b5, b6, b7k1, b8k2.

Total number of unknown variables is 8. Therefore we need 8 equations to derive all values of b and k.

Because,

|m| = |c| = 2

Each input/output pair has 2 equations. That is why we need 8/2 = 4 pairs of  input/output to derive all values of b and k.

-Count

Use BinToImg.py to analyze UEFI BIOS

We can use BinToImg.py to analyze UEFI BIOS. Below is the input files.

• BIOS.rom - A capsule file built by EDK build system.
• SPI0.bin - A binary file dumped from SPI ROM 0 after first booting.
• SPI1.bin - A binary file dumped from SPI ROM 1 after first booting.

The tool BinToImg.py transfers a byte to 8 pixels with black-white color. For example,

Byte = 55h = 01010101

Pixels 
  = 1 -> 0 -> 1 -> 0 -> 1 -> 0 ->1 -> 0

  = black, white, black, white, black, white, black, white

We use the below commands to generate PNG files with black-white bit pixels from the input files.

python3 BinToImg.py -w 6000 BIOS.rom
python3 BinToImg.py -w 6000 SPI0.bin

python3 BinToImg.py -w 6000 SPI1.bin

The option -w 6000 means that the tool generates an image of which width is 6000 pixels.

Below are the generated PNG files. I consider any BIOS engineer know what happens by observing  these images, especially for BIOS.rom.png and SPI1.bin.png.

BIOS.rom.png


SPI1.bin.png


SPI0.bin.png


-Count

Cross Platform Random Bitmap Generator

I created a project, RandomBitmap,  in the below GitHub.
https://github.com/CountChu/RandomBitmap

There are two purposes of the project:

1. to check if random variables generated by a specified platform (e.g., Windows, Android) are secure.
2. to check if a file is random or has patterns by observing the black-white bitmap file transformed by the file.

This idea comes from the page, Pseudo-Random vs. True Random. The author uses PHP to transform random numbers to a bitmap. I separated the transformation into two Python programs, GenRandom.py and BinToImg.py because I want to test random number generators on different platforms as the below picture.

GenRandom, is a set of tools written by different languages on different platforms, calls a specified random generator to produce random numbers. These numbers will be saved in a file. The different languages are specified from the below reasons:

• Python is a cross platform language so that I can use GenRandom.py to test Windows, Linux, and MAC.
• For Android device, Java is a candidate language. Therefore I'll create a Java version tool, GenRandom.java, to test Android's random generator. 
• For security IC or UEFI environment, C is a candidate language. The C version tool is named GenRandom.c.
• For Chrome Browser's Native Client written by C/C++,  the tool will be named GenRandom.c. I'll use it to test random generators of OpenSSL and LIBC in Native Client environment.
• C# is the first candidate language in, I'll have a C# version, GenRandom.cs, to test Windows CNG.

RandomGenerator.py provides different random generators that are consumed by GenRandom.py:

• rg1() is a Python default random generator that is random.randint().
• rg2() is a random generator comes from the page.

BinToImg.py is a python tool that transforms data in a file into an PNG image file with black-white pixels so that we can observe the image to check if these numbers are random. The tool can be used independently. For example, we input any type of file in the tool to generate an picture with black-white pixels to check if the file is randomly.

I use the both tools to verify if the MAC random generator is secure.

>python GenRandom.py Random.bin

>python BinToImg.py Random.bin

I use the tool BinToImg.py to check a PDF file. We can find some patterns in the below picture.

>python BinToImg.py One.pdf

-Count

If PRG is secure, PRG is unpredictable

Below is the definition of secure PRG.




Below is the definition of predictable PRG.



How do we proof the below theorem?

PRG is secure => PRG is unpredictable

The previous page proofed it by:

PRG is predictable => PRG is not secure

This pare proofs it in a direct way. Before proofing it, I  draw the below picture, where G is PRG, to understand the definition of secure PRG and unpredictable PRG.



We keep the picture in our mind and use simple formulas to proof it.

The fact is that G is secure:

Adv = |Pr [A (G) = 1] - Pr [A (r) = 1]| <= e (e is epsilon)

Because the statement, that is proofed by the page, is also true.

Pr [A (r) = 1] = 1/2

Therefore

Pr [A (G) = 1] <= 1/2 + e

We can transfer the statement from A to B for the definition of A that is a statistical test.

Pr [B (G|1...i+1) = G|i+1] <= 1/2 + e

Therefore G is unpredictable.

-Count