Advanced Computing in the Age of AI | Wednesday, April 21, 2021

Intel Xeon Advances Nasdaq’s Homomorphic Encryption R&D 

April 7, 2021 -- Nasdaq, a global technology company serving the capital markets and other industries, is leveraging the crypto acceleration in Intel’s 3rd Generation Intel Xeon Scalable platform to significantly speed up computation for its high performance advanced homomorphic encryption (HE) applications. HE is an emerging form of encryption that allows the ability to compute private data without having to decrypt the data.

Intel continues to introduce new instruction set architecture (ISA) extensions designed to significantly increase cryptographic performance and further enable the emerging HE ecosystem. HE allows users to compute on always-encrypted data. The data never needs to be decrypted, helping reduce the risk from cyberthreats. HE offers developers and enterprises new ways of gaining insights from sensitive data across organizations, and Intel’s platform improves performance to support end-to-end data encryption. To further accelerate HE applications, Intel and Nasdaq are co-engineering HE calculations utilizing the new Advanced Vector Extensions (AVX) 512 Integer Fused Multiply Add Instructions (AVX512_IFMA) available in the latest 3rd Gen Intel Xeon Scalable processor.

To test the real-world applications of HE, Nasdaq validates these capabilities through a variety of proof of concepts. It has been utilizing HE to better enable artificial intelligence and machine learning datasets, as well as apply HE in Nasdaq’s fight against financial crime — particularly around anti-money laundering (AML) and fraud detection. Having the ability to unlock proprietary data without exposing the data itself allows for a stronger collective effort to consolidate unique datasets to analyze and identify potential financial crime activity.

“Homomorphic encryption permits computation on encrypted data without decryption, enabling users to gain new insights from encrypted datasets,” said Nikolai Larbalestier, senior vice president, Enterprise Architecture at Nasdaq. “However, HE is performance-intensive and poses usability challenges for large, enterprise-size datasets. For Nasdaq, we have been exploring and experimenting with HE in strengthening our technology solutions that focus on financial crime detection while also complying with data privacy regulations. The performance improvements achieved with Intel’s latest Xeon platform will support our efforts in wider adoption of HE, particularly in improving data analysis and insights, as well as product innovation around areas such as anti-financial crime.”

Homomorphic encryption (HE) is a powerful new technique for enabling computation and collaboration on private and sensitive data through end-to-end encryption. HE could revolutionize how financial institutions interact with and share datasets for analysis in the future, empowering organizations to gain valuable insights while reducing risk of exposure that could compromise confidentiality.

Trusted execution environments, such as Intel Software Guard Extensions (Intel SGX), also address the protection of data while being processed in memory, though from a slightly different angle. Intel SGX provides a hardware-based trusted enclave for general-purpose workloads to operate within, protecting both data and code from the broader system stack with minimal performance overhead. As performance and usability challenges are resolved, HE promises to bring new options for end-to-end encryption, especially in purpose-specific data-sharing schemes. This is a new frontier and a burgeoning arena for innovation as attacks become more sophisticated and regulatory expectations increase.

Intel also recently announced that it had been selected as a research partner to perform in DARPA’s Data Protection in Virtual Environments (DPRIVE) program. The multi-year program aims to develop an accelerator for fully homomorphic encryption (FHE) to reduce the performance overhead currently associated with it. When realized, the accelerator could deliver a massive improvement in executing FHE workloads. Intel is collaborating with partners to provide leadership in this exciting field, ensuring the broadest and most performant set of options for protecting sensitive data wherever it resides.


Source: Intel

Add a Comment

EnterpriseAI