Compute on all the World's data ...
Imagine being able to do computations on all the world's data, while still protecting the privacy of the information. Data stays encrypted at rest, in transit and with confidential computing, even in use, to comply with the strictest regulations and to enable collaboration without compromise.
Most private and public organisations today operate on data that should or are required to stay confidential. At Partisia we provide the Partisia Confidential Computing platform, which can guarantee that all computations on any confidential data are kept highly secure, auditable and most importantly confidential and still be able to provide output, which can be used anywhere in the world.
Let's talk about Confidential Computing - Full video 2:39 mins
Case Study
Virtual Public Register
The Virtual Public Register (VPR) Platform represents a system that facilitates confidential computing and collaboration across public registers i.e. a system that virtually combines and uses public registers.
The VPR Platform has been matured for commercial use through a series of R&D projects funded by the Danish Industry Foundation and Innovation Fund Denmark. These projects focus on healthcare and are used within a “sandbox” environment where the Danish Health Data Organisation (SDS) and Statistics Denmark (DST) operate and control the VPR Platform (they are the Computing Parties running the MPC protocols). The MPC protocol ensures that neither of the public registers can access the encrypted data. The user interface allows third party analysts to select data sets, variables and computations across the merged data sets. If the SDS and SDT approve the analysis, the third-party analyst receives the results.
The OSCAR Project aims at building a commercial platform around the VPR Platform and where a number of template healthcare analyses are readily available for third-party analysts. The potential is a real-time access to run analysis on highly sensitive data that is very time consuming and expensive to apply today.
Product decription
The platform enables confidential data collaboration among organisations without compromise.
Partisia Confidential Computing platform provides an easy programmable solution to doing computation on encrypted data. Confidential Computing facilitates large private and public data providers to collaborate around activating confidential data using a combination of MPC and Federated or local computations. The UI provides an easy way to connect to external data sources and Business Intelligence (BI) tools. The blockchain based orchestration ensures a transparent and trustworthy collaboration between the data providers and other users of the system.
Partisia Confidential Computing supports both realtime and batch analysis. Depending on the complexity of the data and the required confidential computations, both realtime and batch analysis can be activated through any integrated BI tool, calling a deployed analysis on the platform.
As a standard enterprise solution, the platform supports integration with any authentication and authorization system, so organizations can define who is allowed to run confidential computations on specific data.
All data can be computed on directly in encrypted form
Full auditability of all use, making regulatory compliance seamless
Computations are defined by using either Rust, Java or bespoked frontend
Integrations with most data sources and BI tools
Authentication and authorization integrations
Deployable both on-premise and in cloud
Due to regulation, data must be encrypted at all states (At rest, in transit and in use)
Due to a lack of security procedures and human errors, breaches of personal data happen daily.
Reduce costly procedures for data usage approvals and speed up the flow around the use of sensitive data.
Data is lying stale in public repositories, given trust and security, they can bring value commercially as well as for the public good
Collaborations between corporations with sensitive data is almost impossible today, since data needs to stay protected and encrypted.
As an industry, we should innovate services through data collaboration, but lack of insight, confidentiality concerns and regulations prevent industry wide data collaborations.
Case
Financial Fraud Detection
Financial fraud is an increasing problem that globally costs trillions for dollars and fuels organized crime from drug dealing to human trafficking. The fraudsters become more and more advanced and money moves around in patterns that become harder and harder to distinguish from normal behaviour.
Fortunately, most financial fraud passes through banks and other financial institutions, which leave digital traces originating from single transactions between a sender and a receiver. Most fraud detection today is conducted within a single bank with too little information to recognise fraud. Information has to be shared across not just the sender and the receiver but the large number of banks involved in the chain of transactions used to disguise fraud.
This is, however, where Confidential Computing can play an important role. Done the right way, the Partisia Platform and MPC allows the use of confidential data across banks, regulators and jurisdictions without violating either regulation that ensures competition and the individuals’ right to privacy. This has been recognised by regulators and recently, FCA (the financial regulator in England) hosted an event where Partisia developed a PoC solution in close collaboration with Goldman Sachs, Deloitte, Partisia, Sedicii and Ex Ante Advisory.
Use Cases
Health and Pharma
Valuable well-structured data collected over decades by health authorities and national statistics agencies can now be shared without revealing it. Pharma companies and national drug procurement can now get insights on both drug development, usage and effects much faster and cheaper. In Denmark ongoing R&D projects aim at utilizing this to turn Denmark into one big privacy-preserved and secure phase 4 drug study.
Fraud Detection
Financial institutions can now perform risk assessments and fraud detection on their confidential data, this can even be extended with multiple institutions sharing data fully confidentially, without any party learning anything about any other party.
This enables multiple data silos to be merged together, without ever breaching the confidentiality of the information.
Training AI
Combining MPC and Federate Machine Learning enables faster privacy-preserving training of AI models. A structured and transparent orchestration ensures that the most computational intensive training happens distributed on the data providers own data. The insights across the local training is combined with MPC in a privacy-preserving way. Collectively these two Privacy Enhancing Technologies pave the way for activating better and more sensitive data in training AI models for better decision making across applications.
Enabling independent organisations to work privacy-preserving on joint data without regulatory or antitrust challenges is a game changer.
Documentation
How to deploy?
Scenario 1:
Customer deploy and operate the network as a single entity: The customer downloads the software and deploys with a single click a network controlled entirely by the customer …
Scenario 2:
Customer join an operational network as node operator: The customer gets invited to the existing network and downloads the software deployed with a single click …
Scenario 3:
Customer join an operational network: Here the starting point is that there already exists a network of node operators that run the Partisia Platform and Confidential Computing. This reduces the integration to simply integrate the APIs or build tailored services using the built-in smart contract language.
Our products
Solutions to activate sensitive data anywhere
