News blog

Some holiday reading on cool new Sharemind features

It's that time of the year when we can look back on what was and, more importantly, think on how we can learn from it for the future.

We have been hard at work on the next generation of secure computation tools and we know that you have waited for a while already. So here's some peeks into what's coming. There are two preprints of papers for your holiday reading pleasure.

Kamm, Liina., Willemson, Jan. Secure Floating-Point Arithmetic and Private Satellite Collision Analysis. Cryptology ePrint Archive, Report 2013/850. 2013. (PDF)

Bogdanov, Dan., Kamm, Liina., Laur, Sven., Pruulmann-Vengerfeldt, Pille. Secure multi-party data analysis: end user validation and practical experiments. Cryptology ePrint Archive, Report 2013/826. 2013. (PDF)

The two papers show how Sharemind has grown from working just with integers to solving complex data analysis problems.

We wish you a peaceful end of the year and are looking forward to sharing great news with you in 2014!

--- the Sharemind team

Sharemind powers a new cloud-based data analysis service

Sharing data with a cloud service means trusting the service provider with your information. While there are many cases where the standard security guarantees are sufficient, medical data, financial data and personal data often requires better privacy.

Secure multi-party computation can provide that privacy and now, with Sharemind on the Cloud, we can do it on standard cloud servers.

Our new Income analysis of the Estonian public sector cloud application is powered by Sharemind and runs on Amazon EC2, Microsoft Azure and Estonia's own Zone Media cloud. No cloud service provider among these three can access the individual incomes or find out who they belonged to. Sharemind can still calculate the reports and present it as a nice web service.

Click here for the demo!

We will continue working on improving the security of cloud services with Sharemind.

A glimpse of the new SecreC language

We have been dropping hints of massive updates to the Sharemind SDK. While we are still quite a bit away from releasing the tools, we are publishing bits of information about the upcoming changes.

One of the most import changes is descibed the following ePrint:

Bogdanov, Dan., Laud, Peeter., Randmets, Jaak. Domain-Polymorphic Programming of Privacy-Preserving Applications. Cryptology ePrint Archive, Report 2013/371. 2013. (PDF)

The new SecreC 2 language and the Sharemind 3 server will be based on protection domains. A protection domain is a set of secure data representations and algorithms for secure computation on that representation. You can think of a protection domain as a set of secure computation protocols.

This allows us to start using different secure computation protocol suites in Sharemind with little effort. Sharemind 3 already supports various protection domains with a different number of servers.

We will be posting more information about Sharemind 3 as it becomes available. Stay tuned!

Sharemind SDK v2013.05 released

After a long development period, we are very happy to announce the release of Sharemind SDK 2013.05.

The highlights of this release are:

  • a completely redesigned secure database
  • a new standalone virtual machine package for Linux, Mac OS X and Solaris users
  • new tutorials, updated documentation and examples.

Get the new Sharemind SDK 2013.05 here

Check out the new tutorials!

  1. Tutorial: How to install the Sharemind SDK virtual machine image with VirtualBox
  2. Tutorial: Understanding the secret-shared database in Sharemind
  3. Tutorial: How to collect and use data in the Sharemind database

Check out the new database API

SecreC Database instructions for Sharemind 2.3

Here is the full change log:

Release notes for SDK 2013.05 / Sharemind 2.3

General

  • Added persistent sessions: controller applications are now able to disconnect without losing their session.
  • Redesigned the Sharemind database architecture: the database is now more flexible.
  • Added the HDF5 database backend (enabled by default).
  • Refactored the Tokyo Cabinet database backend.
  • Removed the SQLite and ODBC database backends.
  • Removed the database related functions from the ControllerInterface.
  • Updated the unit tests to cover the new database functionality.
  • Added a system call mechanism for more easily adding new protocols implemented in C++.
  • Added system calls for sorting vectors and matrices. These are optimized for specific input sizes.

Sharemind languages

  • Added and updated a number of functions in SecreC in relation to various changes (e.g. database functions).
  • Added and updated a number of functions in Sharemind assembly in relation to various changes (e.g. database functions).

SecreCIDE

  • Made SecreCIDE compatible with Qt5.
  • Updated the Sharemind assembly language reference in the help section to take into account the changes in the language.
  • Updated the SecreC language reference in the help section to take into account the changes in the language.
  • Updated the syntax highlighting in the code editor to account for the changes in the Sharemind languages.

Bug fixes

  • Fixed some SecreC compiler bugs.
  • Fixed some unit tests.
  • Fixed an issue where some mutexes were not properly locked.
  • Other various bugfixes.

Cybernetica develops a solution for increasing space security

Cybernetica has developed a system the lets satellite owners find the probability that their satellites collide with other objects in the Earth's orbit. Our solution is revolutionary, as it protects the confidentiality of the trajectories - the owners do not need to share the locations with the satellite with other parties.

The solution is built on the Sharemind 3 secure database and application server that protects data from unauthorized access using cryptographic techniques. The use of secure multiparty computation ensures that even the organization that hosts the Sharemind system is unable to learn the secret values. Sharemind 3 is the first practical secure multiparty computation system that implements programmable secure computations on precise floating point numbers.

The solution has been developed for the United States DARPA PROCEED programme and was successfully demonstrated earlier this year. Cybernetica will continue to develop solutions to increase the security of data sharing for all applications.

Contact us for more information.

A PhD thesis describing the full Sharemind model is now available

If you ever wanted to get a complete overview of the Sharemind system - the fundamentals, the protocols, performance, programming and applications, now there is a great place for that. The Sharemind database and application model is now described in a doctoral thesis, defended by Dan Bogdanov on February 28th, 2013 in the University of Tartu.

Dan Bogdanov. Sharemind: programmable secure computations with practical applications. PhD thesis, University of Tartu, 2013.

PhD examiners: Nigel P. Smart (University of Bristol) and Berry Schoenmakers (TU Eindhoven)

Full version in PDF

Video of the defence

If you need citing information, see the Research section.

Syndicate content