What is AKDB?

So, what is it?

AK (Russian: Автомат Калашникова, English: Avtomat Kalasnikova) DB is a simple relational database management system developed by professor Markus Schatten of Faculty of Organization and Informatics in Varaždin, Croatia. It is also regularly developed, maintained and fixed by graduate students attending 'Physical Database Modeling' course every spring.

The development started in 2009 and is still ongoing, as building a custom DBMS is not an easy task. Most of it is written in C, with some Python for Server stuff and documentation which is generated automatically and can be viewed in HTML or LaTeX form. More on that in Installation and Build pages.

Where can I find it?

The AKDB System can be found at GitHub, it is maintained there for the last couple of years. Before Github, AKDB System was maintained on Sourceforge.

It is also important to mention that the system is currently in ALPHA development and is NOT fully functional!

Download from GitHub



Originally built for Unix based operating systems, AKDB has a pretty straightforward installation on these systems. To get your AKDB going you need to simply clone the repository and install dependencies, but first you need to install prerequisites:

sudo apt-get install git build-essential swig make doxygen

To get Kalasnikov DB you can download .zip file from GitHub on your machine or you can clone it into a directory you've positioned yourself in by using Git command:

git clone https://github.com/mschatten/akdb

You can then change the directory to get into the structure:

cd akdb/akdb/

Nevermind the multiple akdb directories, it was probably just a bug during migration to GitHub. In the end you should end up in the root directory of the system.

Next thing you do is install dependencies (Python 2.7), if you don't have them already:

sudo apt-get install python2.7

You can also set up a virtual environment for the system. To do that, install virtualenv via pip and run

virtualenv -p /usr/bin/python2.7 venv

into your project folder. You can do this with any Python interpreter of your choice.

To use it, this environment needs to be activated with source venv/bin/activate and deactivated with command deactivate.


[UNTESTED] Everything from the Linux instructions should work. Please let us know on GitHub if it doesn't.


Windows 10 introduced Windows Subsystem for Linux (WSL) in 2016, which enables running instances of Linux distributions on Windows. Running instances can use their file system (hosted on the Windows partition) or they can access existing Windows files, that can make some things easier.

Enabling and configuring WSL


Hover over images to to zoom in.

Step 1

Click on the search button (or search box if it's enabled) or press the Windows key and start typing.

Search for turn windows, you should see and click on Turn Windows features on or off.

Search box

Step 2

In the list that appears scroll down until you find Windows Subsystem for Linux.

Check it and click OK. Wait for the feature to be enabled and after it finishes restart your computer.

Search box

Step 3

Open the Microsoft Store app. In the search box type linux. Select Run Linux on Windows.

Pick a distribution and click Get. Wait for everything to install (download should be about 200-250MB).

Search box Search box Search box

Step 4

Open PowerShell as an administrator. Easiest way is to right-click on the Windows button in the bottom-left, and then clicking on Windows PowerShell (Admin).

Type in a command to initialize the WSL installation (depends on the distribution, check the Store listing), eg. ubuntu for Ubuntu or opensuse-42 for openSUSE.

You can now run Linux on your Windows device! Hooray!

Search box Search box

Step 5

Follow the guide for Linux.

You might need to install some dependencies like LaTeX for everything to work without warnings or errors.

WSL tip

After you close the PowerShell window, you can bring back the Linux terminal by typing bash in the search bar, the Run dialog or by Shift-right clicking in File Explorer then clicking on Open Linux shell here, which will launch a bash shell with the location set to the current folder.

Search box

Docker (Windows, Mac, Linux)

You can get a Docker image that enables you to run AKDB inside Docker container without the need for a Linux virtual machine or WSL.

Requirements for this are Docker and Make, which you have to install on your local machine. All other prerequisites will be downloaded and installed automatically.

You also need to install Docker container and this will take some time, but it allows you to develop the application in your native environment and then compile and test its code inside the container. For building AKDB in Docker container please visit Build section.

It is also required of Windows users to turn on drive sharing in Docker settings. To do that open Docker settings, go to shared drives, check the drive on which AKDB is located and click Apply.

Get Docker
Get Make



After you've installed everything system needs to run you can now build the application. To do this you need to position yourself into the src directory and execute:

make doc
make swig

This will compile and build the application, along with creating documentation in Doxygen and connecting C with Python using SWIG. To view documentation go to the doc directory and open a PDF document called refman.pdf.

To run AKDB position yourself into the bin directory and execute:


This will initialize disk manager and prepare all tests. To see additional commands type:

./akdb help

You can see other currently available options are testing options, and will be explained in the testing section. If you want to clean the project of files generated during compiling run this:

make clean
make clean-d

To see if there are some leftover Croatian comments in the code:

make comments

As it was said before, the application is not yet functional, so this is currently used for developing and testing purposes.


After you've installed Docker and Make you can start the Docker container with AKDB and its host directory will be automatically mapped in the container directory. This means that all the changes you make on your machine are automatically seen and saved inside the container.

For more information examine the Dockerfile in the root directory. To install Docker execute command:

make docker-install

If you're running on Windows enter:

make docker-win

If you're running on Mac enter:

make docker-mac

All other commands for starting the application and performing tests are the same like in a native environment.


AKDB system root directory consists of three folders and two files. Files contained within this directory are LICENSE file (contains the GNU General Public Licence) and a README file which in short describes installation and building of the application. Folders in the root directory and their purpose be explained next.

Binaries (bin)

Essential folder where the application compiles to and from where you can run it. It contains several files and another folder called blobs where all binary files are stored after running. The blob folder is created empty, and is created after the application is installed and started for the first time. Files in the bin folder are:

Documentation (doc)

Documentation for AKDB is created using Doxygen. It's a tool that generates documentation from comments in the source code. The end result is a reference manual in PDF format. Doxygen can create this in HTML or LaTeX format.

In this folder there are two additional folders. The HTML folder contains the HTML and JS elements used for creation of the web version of the manual while the LaTeX folder contains TeX elements for creating PDF and RTF manuals.

A lengthy AKDB reference manual in the PDF format called refman is located in the doc folder. It's highly recommended that you study it if you want to get a better insight into the system.

Source (src)

This folder containts the main source code for compiling and building the application. It has files required to start building process and subfolders where the source code files are located.

Most of the source code files are written in C, expect Server and SWIG files which are written in Python and shell scripts.

Files in source directory are:

Folder Description
Auxiliary (auxi) Debugger, Type converter, Constants, Dictionary, Parser, Memory Programmer, Observer, Test reporting
Disk manager (dm) Source code for managing database and its objects on hard disk
File manager (file) Database indexes (bitmap, btree, hash, regular), blob management, File input and output, File search, File sorting, File identification, Table managment, Sequences, File manager testing
HTML (html_files) Header and footer for Doxygen in HTML (includes CSS and JS scripts)
Memory Manager (mm) Utility that manages memory allocation and deallocation in database operation
Optimization (opti) Query optimization, Relational operations optimization (association, comutation, projection, selection)
Recovery (rec) Archive and redo log manager, Recovery tools for system crash, Database log in textual and binary form
Relational algebra (rel) Operations that are used in queries and are essential for database functionality (aggregation, difference, expression check, intersect, natural Join, product, projection, selection, theta join, union)
Server (sql) Client and Server source code (running and configuration), SQL executor
SWIG (sql) Wrapper files for integration of C with Python in SWIG tool (includes configuration and test)
Shell tools (tools) Generation comments, file fetching, C and Pythin parsing, application updating
Transactions (trans) Source code for performing transactions in the database


Since the project is still in development, tests are required to ensure its stability and functionality. There are couple of dozen tests that check whether certain parts of the application work correctly.

After you've built the application, to run all tests simultaneously enter the command:

./akdb alltest

On first use DB file will be initialized. In case it hangs press 1 and Enter (don't ask why). It also is not unusual for program to crash while running certain tests, because there are still some unfixed bugs that can cause system failure, but most tests work.

To view what current tests are available, run the following:

./akdb test show

This will write out the list of current tests that can be performed. As of 21st of March 2018 they are numbered and divided into few categories:




3Observable pattern

Disk manager

4Allocation bit

5Allocation table

6Thread safe block

File system




10File IO

11OP rename

12File sorting

13File search


15OP table



17Binary tree


Memory manager

19Memory manager

20Block memory


21RE association

22RE commutation

23RE selection

24RE projection

25Query optimization

Relational operators







32Selection pattern

33Selection (redo)

34Check expression



37Theta join

SQL tests













50Constraint names






Crashes the application

Performing this test twice in a row ruins the application

Takes forever to perform and you have to kill the app

Show tests command won't enable you to perform those tests inside the application. For that, you need to use:

./akdb test

This will create the necessary test tables on which tests are run. To run a test enter a number of test you wish to run. If you want to run a particular test you can enter the following command in shell:

./akdb test [number]

This will automatically execute the entered test and exit the application.

Client and server tests

In order to start these test you must position yourself in the src/srv directory and execute test commands by shell.

Server test:

python server.py

Client test:

python akdb_cli.py

It is also recommended to read the README file in the src/srv directory to get a more detailed overview of these tests.

Known issues

Issues that are being worked on are listed on GitHub and are generally updated every year. They are mostly in Croatian.

You can contact prof. Markus Schatten if you have any questions or want to volunteer for work on AKDB.

GitHub issue page
Contact prof. Schatten