Paul Wee, Head of Product, IXUP
With the Big Data revolution came the realisation that connecting data from different divisions and organisations produces richer insights. This potential reward, however, wasn’t without significant risk and it became evident that widely used data sharing practices weren’t secure or sustainable.
Traditional data sharing practices
Data collaboration between parties is generally performed through legal contracts or the use of trusted third parties. These can be slow and costly, and ultimately requires unencrypted data to be handed over, with the ever-present threat of data loss and misuse.
The rise of homomorphic encryption
Enter the era of encrypted data collaboration, where companies no longer need to make the trade-off between data utility and security.
Through advances in cryptographic research, namely homomorphic encryption, organisations can now perform calculations on multiple data sets without ever needing to decrypt them. Unlike traditional techniques, homomorphic encryption protects data at every step – at rest, in transit, and in use – so at no point is the data vulnerable.
Due to performance reasons, homomorphic encryption has been confined mostly to academia for decades. However, as a result of advances in research and computational power, what was once a largely theoretical area of cryptography is now viable for commercial use. In a future blog post, we look forward to sharing with you the performance capability of our homomorphic encryption engine.
Top considerations when choosing a data collaboration platform
- Encrypted-by-design: Look for a system where encryption is not just an afterthought – an encrypted-by-design philosophy should underpin every design decision. From the overall architecture down to the smallest feature, the IXUP team is constantly challenging each other with the question: how can we make this work on encrypted data?
- Tried and tested: Look for a system which is very selective about where it attempts to reinvent the wheel. While it can be exciting to create something new, going against the grain of decades of cryptographic and data security research can be fraught with danger. IXUP’s philosophy is to “stand on the shoulders of giants” – we’re incorporating the best and most globally recognised processes and technologies, such as the Five Safes governance framework and Microsoft’s SEAL library for our homomorphic computations. When it comes to your company’s reputation, you want to rely on processes and technologies which have been subject to years of industry and academic scrutiny and refinement.
- Control: Look for a system where you are always in complete control. On the IXUP platform, data encryption and decryption is performed exclusively client-side, and you hold the encryption keys at all times; they never enter the IXUP platform. Also, our Design Canvas – our ETL modelling tool – provides you with a clear and unambiguous view of how data is being used, while being provided with field-level controls for modelling, analysing, and deleting your data.
- Ease of use: Look for a system which is easy to use. A research paper by Desai and Ritchie (2010) noted one of the primary causes of data breaches was inconvenience, not malicious intent. Basically, if it’s too hard to use, even the most secure system in the world may fail to protect your data. At IXUP, the user experience is always placed at the centre of the design process, which led to the creation of our Design Canvas. It features a highly visual and easy-to-use drag-and-drop interface, empowering data analysts of any skill level with some of the most sophisticated encryption technologies the industry has to offer.
- Multi-dimensional governance: Look for a system which goes beyond the data itself and relies on more than just an audit log. While data on the IXUP platform remains encrypted at all times, we took this a step further by adopting a multi-dimensional approach to data security. Using the Five Safes governance framework, in addition to the data itself, we also incorporate the project’s objectives, the people involved, the settings and the outputs as part of the collaboration approval process.
Data collaboration: reward without the risk
It’s an exciting time for truly secure data collaboration between multiple parties, with technological advancements forging a more accessible, commercially viable path which in turn has lowered the barrier of entry for organisations. I’m proud of the product we are building here at IXUP. We are constantly evolving it to capitalise on technological advancements, responding to the ever-increasing risks in the environment and, most importantly, the needs of organisations to make the most of their data.
Computational power is only going to continue to grow, especially as Artificial Intelligence (AI) and machine learning technologies continue to evolve. The implications for better services across industries and advances in science and environmental modelling, for example, are profound.
The tips outlined in this article can help guide organisations to select an expert to partner with for their secure data collaboration needs.