EasySMPC is an app for securely summing up distributed confidential data using Secure Multi-Party Computation (SMPC). It is designed to enable simple statistical analysis with maximum usability, easy access and a strict no-coding approach. Your parents should be able to use this and so should your physician!
Prerequisites ⚫ Installation ⚫ Features ⚫ Screenshot ⚫ Quick start guide ⚫ Contact ⚫ License ⚫ Acknowledgements
EasySMPC requires Java, at least in version 13. The Java runtime is bundled in our Installer package.
To compile the app from source in addition to the Java JDK the Maven build system is required.
EasySMPC does not need an installation and can be used as a Java jar package. However, to increase portability we packaged the nessecary Java runtime with our application in an installer to build an executable for Linux, Windows and Mac OSX. This installer does not need any administrator priviledges and should be run as a user. The installers for Windows and MacOS are not signed. Thus a respective message while installing must be confirmed.
Check out our releases page for Windows, Linux and MacOS executables.
To build the executable yourself, please clone this repository and build with
maven (mvn package). The assembled executable should be in the target
directory. At present time some tests occasionaly fail. We're looking into that.
Until those tests are passing please compile with mvn package -DskipTests.
To build the installer please build the jar package as described above and then use the supplied scripts for your target platform. E.g.:
cd installer && ./linux.sh
EasySMPC was built to allow non-technical personell in medical research perform simple analysis without sharing their input data. We tried to achieve a very low threshold of technical prerequisites by using email as an, in most cases, already established and configured communication medium.
- Easy to use
- Communication using established channels, e.g. emails
- Excel and CSV import and export
- Automation of the protocol using a shared IMAP-Mailbox
- Automatic Proxy-Detection
EasySMPC uses Arithmetic Secret Sharing
[DZS15],
the arithmetic extension of the GMW-Protocol
[GMW87] to achieve the private
computation of the sums. It uses a ring of size , the 12th Mersenne
prime.
- Differential Privacy
- Automate email communication with private mailboxes
- Support for decimal numbers
- Use EasySMPC with Slack/Mattermost/IRC/...
- Examples and Getting Started guides
- Further Documentation
-
As a study initiator, click on Create new project and enter the names of all participants, their e-mail addresses, the variables to sum up and your own confidential data. For performing the computation, you can choose between an automatic or a manual mode.
a) In manual mode the users need to exchange all messages by sending and receiving e-mails manually.
b) In automatic mode participants receive and import the initial message manually into EasySMPC (see 2). All further messages are exchanged automatically via their respective mailboxes. If the study initiator chooses automatic mode, all participants also have to choose automated mode. The user name and connection details provided for the automated mode will be saved for future use, the password will not be saved and must be re-entered if the study was closed in between.
-
As a participant, you copy the message you received via email into the clipboard, click on Participate in project in EasySMPC and paste the content. You will now see the study definition and can enter your own confidential data and mail box details if applicable.
-
As an initiator or participant, you now click on proceed. If running in automated mode, EasySMPC will automatically perform all steps until the final result is displayed. If running in manual mode, all users need to send and receive e-mails prepared by EasySMPC to perform the computation.
-
The final perspective shows the result of the secure addition of all variables.
If you have questions or problems, we would like to invite you to open an issue at Github. This allows other users to collaborate and (hopefully) answer your question in a timely manner. If your request contains confidential information or is not suited for a public issue, send us an email.
EasySMPC's core development team consists of:
- Tobias Kussel - Cryptography - tobias.kussel@cysec.de
- Armin Müller - Graphical user interface - armin.mueller@charite.de
- Fabian Prasser - Software design and architecture -fabian.prasser@charite.de
- Felix Nikolaus Wirth - Graphical user interface - felix-nikolaus.wirth@charite.de
This software is licensed under the Apache License 2.0. The full text is accessible in the LICENSE file.
EasySMPC uses the following dependencies:
- FlatLaf - Apache License 2.0
- Apache HTTPClient - Apache License 2.0
- HPPC: High Performance Primitive Collections - Apache License 2.0
- Apache POI-OOXML - Apache License 2.0
- Apache Commons CSV - Apache License 2.0
- jUnit (Unit tests only) - Eclipse Public License v2.0
- Proxy Vole - Apache License 2.0
- Apache Logging - Apache License 2.0
- LMAX Disruptor - Apache License 2.0
This project is partly financed by the "Collaboration on Rare Diseases" of the Medical Informatics Initiative, funded by the German Federal Ministry of Education and Research (BMBF).