#+STARTUP: inlineimages

#+BIBLIOGRAPHY: refs.bib

LibQrng

Introduction

Through the software solution presented in this paper, we provide a method to interact with IDQ's Quantis Appliance network-attached device---a quantum random number generator that securely generates high-quality random numbers [@noauthor_quantis_nodate]---through REST API. We deliver an easy-to-extend solution for various programming languages and integrated projects. As such, our solution's implementation supports the straightforward use of quantum-generated random numbers in applications and research areas in which the output of the experiments is highly dependent on the quality of random numbers.

In the current form, libqrng offers access to IDQ's Quantis Appliance through an API that can be easily integrated into applications that need true random numbers. The library is written in C language and is currently available as a library only on GNU/Linux-based operating systems. We decided to implement it as a library in native C since, to the best of our knowledge, it can be easily used with other programming languages via Foreign Function Interface (FFI). As a proof of concept, we implement Python bindings to libqrng and provide examples of how to request random numbers from IDQ's device.

We designed the software stack we provide to be easily extended to different scientific programming languages and scientific/engineering fields, such as cryptography[@bruce1996applied], Monte-Carlo methods[@ferrenberg1992monte], heuristic algorithms[@navarro2022review], industrial testing and labeling, hazard games, [@stipvcevic2011quantum], etc.

Software Architecture

IDQ's Quantis Appliance device is connected to the Local Area Network so that the random bytes are retrieved using REST APIs via HTTPS protocol. As we show in Figure fig:archblock, our proposed solution for using the IDQ's Quantis Appliance devices comprises 3 software layers: the libqrng library (which uses libcurl to request data from the device by GET requests), the Foreign Function Interface used for calling the library API from different programming languages (currently we support the Python programming language by utilizing the Python extension), and the Application Layer (for which we provide application examples that use the library directly or through the Foreign Function Interface).

Our libqrng is a shared object that exports APIs to request random bytes from the IDQ Quantis Appliance. Figure fig:libqrng_init presents the library's initialization sequence. As shown, the library uses libcurl to perform network requests. The device_domain_address parameter is mandatory since it represents the IP address of the quantum random generator device. Figures fig:libqrng_stream, fig:libqrng_double, and fig:libqrng_int present the sequences that retrieve the random bytes from the device and perform an interaction with libcurl. The API to get a stream of random bytes, as shown in Figure fig:libqrng_stream, is qrng_random_stream and accepts 2 parameters: the stream buffer that stores the bytes, and the size of the requested stream. The APIs for getting random integers (32-bit or 64-bit integers) or doubles (float or double values), as presented in Figures fig:libqrng_double and fig:libqrng_int, accept 4 parameters: min (inclusive) and max (exclusive) that represent the number's range, samples that represent the number of requested samples, and buffer that stores the values. Figure fig:libqrng_close presents the cleanup sequence.

![Sequence diagram presenting the option to get a random double value in a specified range $min,max)$.

![Sequence diagram presenting the option to get a random integer value on 64 bits in a specified range $min,max)$.

Getting started

Prerequisites

1. Operating System

   • Linux based distribution

2. Dependencies

   • gcc (GNU C Compiler)
   • libc (GNU C Library)
   • Make
   • libcurl-dev
   • Python

Installation

1. Clone the repository or download the code.

   git clone https://github.com/sebastianardelean/libqrng.git

2. Change directory to src.

   cd src

3. Build and install the libqrng library.

   make && make install

The make script will build the library and create the shared object. The install target will copy the shared object libqrng.so.1.0 to /usr/local/lib and will create the symlinks to /usr/local/lib/libqrng.so.1.0 and /usr/local/lib/libqrng.so.

The Python FFI is not built and installed by default. To use the Python bindings, first build and install succcessfully libqrng following the above mentioned steps. Then, the next steps must be followed:

1. Change directory to bindings/python.

   cd bindings/python

2. Build and install.

   make && make install

Another option to install the libqrng library is to download the amd64 binaries provided as deb packages (only if you're using Debian based Linux distributions) or as tar.gz archives. The deb package can be installed using dpkg -i package_name.

If you choose to use the tar.gz archive:

1. Extract the content of the archive. tar -xvf archive_name

2. Make the install script executable. chmod +x install.sh

3. Run the install script. The install script will copy the shared object to /usr/lib/, create the symlinks and run ldconfig. ./install.sh

Note The amd64 binaries provided for download do not include the Python binding library!

Examples

In examples directory we provide examples of using each API of the library. As such, we provide the following examples:

1. fwversion to get the firmware version.
2. randbytestream to request streams of random bytes.
3. randdouble to request random double values.
4. randfloat to request random float values.
5. randint32 to request random 32-bit integer values.
6. randint64 to request random 64-bit integer values.
7. sysinfo to request the system informations.
8. knapsack-ga presents the use of the Python bindings and libqrng to implement the genetic algorithm for solving the knapsack problem.
9. simulated-annealing presents the use of the Python bindings and libqrng to implement the simulated annealing.
10. qrand is a command-line tool that uses all the capabilities of the libqrng to request random values.

#+CITE_EXPORT: csl ~/.emacs.d/ieee.csl

#+PRINT_BIBLIOGRAPHY: