For companies in road transport that means, although in theory they get nice data about their drivers and vehicles, in practice they have to pay a lot of money for the digital tachograph and the associated equipment, which is then used against them – the old fraud schemes no longer work, the machine cruelly gives every police officer the driving times of the last 28 days (and could give much more).

The usability of the devices I’m familiar with is also abysmal. It just takes forever to read out the data (9600 baud per default, 115000 baud maximal, but i guess the company’s software does just the former), and the drivers have to keep track of their times manually – just not up to par with 2009 technology.

The company also had just a very bad software for analyzing the data recorded by the card and the digital tachograph, so I was asked to write a simple visualization program.

Compiling

You can get the program files from this page, or from the development page at Sourceforge.
Since so far there is no binary distribution for the program, you will have to compile it yourself.
To do that, you need a c++ compiler, parts of the boost library (specifically program_options and shared_ptr) and the Gnu MP library, wich is used to check the various RSA signatures. Once all those are installed, typing make/make install should work fine. On Ubuntu systems, you will need to do something like:

svn co https://readesm.svn.sourceforge.net/svnroot/readesm readesm
sudo apt-get install libboost-program-options-dev libgmpxx4ldbl  libgmp3-dev
make
sudo make install

You can of course also use checkinstall instead of make install, or type make package, which invokes checkinstall.

Running the program

In most circumstances, you will run the program from the commandline like this:

readesm --infile foo.esm --outfile bar.html --format=html

Alternatively, if no output file is specified, stdout is used. For KDE users, i wrote a little wrapper script named readesm-wrap-kde.sh. It will get installed my typing make install

Security

It is a really pleasant surprise to see a nice security model in the digital tachographs, considering the rules for the security implementations were made by politicians. Both cards and vehicle units work with 1024-bit RSA keys, and each vehicle unit has its own key, with an certificate signed by the member state, which in turn is signed by an European master key. Data is hashed with SHA-1, subsequently padded and signed by the vehicle unit, and that signature appended to the readouts. The law even states that the companies have to store the data in this signed form, so there is little chance to tamper with the data, once recorded.

Since the most likely attacker – the evil manager who wants to exploit the drivers – has physical access to the card and vehicle unit, which contain the private keys, even 1024-bit RSA provides no absolute security. The manager could try a timing attack, or take a really close look at the storage, both times avoiding having to solve the RSA problem.

The connection from tachograph to sensor is also secured, using Triple-DES. The week point here is the sensor – a successful attack against the sensors some DAF trucks are equipped with is to disturb it using strong permanent magnets, thereby preventing it from recognizing the changing magnetic fields.

All possible attacks against the security system however face the problem of being unveiled by police checkpoints or a cross-correlation of the faked data with data from other sources, for example the toll systems recording every few highway kilometers in Germany.

So, although I could sell manipulation programs for quite some money, all I can offer is a program that can detect such attacks

 

readesm.tar.bz2