March 8, 2021 – Intel announced today that it has signed an agreement with the Defense Advanced Research Projects Agency (DARPA) to deliver in the Data Protection in Virtual Environments (DPRIVE) program. The program aims to develop an accelerator for fully homomorphic (FHE) encryption. Microsoft is the leading cloud ecosystem and homomorphic encryption partner that is leading the commercial adoption of the technology once it has been developed by testing it in its cloud offerings, bringing its Microsoft Azure and Microsoft JEDI cloud, owned by the US government. The multi-tool program represents a cross-team effort across several Intel organizations, including Intel Labs, the Design Engineering Group and the Data Platforms Group, to address the “ultimate end” in data privacy, which computes fully encrypted data without the possibility of decryption. keys.
“Completely homomorphic encryption remains the holy grail in the effort to keep data secure while in use. Despite strong advances in reliable execution environments and other confidential computing technologies to protect data while at rest and in motion, data is decrypted during computing, opening up the potential for attacks. which may be present at this stage. This often hinders our ability to share and extract maximum value from data. We are delighted to have been selected as a technology partner by DARPA and look forward to working with them as well as Microsoft to advance this next chapter in Secret Computing and the promise of full homomorphic encryption for everyone. ”
– Rosario Cammarota, chief engineer, Intel Labs, and chief analyst, DARPA DPRIVE program
Why it’s important to: Protecting the confidentiality of vital information – whether personal data or corporate intellectual property – is of paramount importance to businesses. Today, many rely on a number of data encryption methods to protect information while it is being transferred, used and at rest. However, these methods require data to be decrypted for processing. It is through this encrypted state that data can be more vulnerable to misuse.
Fully homomorphic encryption allows users to account for ever-encrypted data, or cryptograms. It will never be necessary to decrypt the data, reducing the potential for cyberthreats. FHE, when implemented at scale, would allow organizations to use methods, such as machine learning, to extract full value from large datasets while protecting data confidentiality across the data lifecycle. . Customers across businesses such as healthcare, insurance and finance would benefit from new practices made possible by being able to maximize and extract value from sensitive data without risk of being seen.
About Democracy adopting completely homomorphic encryption: FHE adoption in the industry has been slow as data processing using fully homomorphic encryption methods on data-intensive cryptograms and imposes a large “performance fee” even for simple operations.
Under the DARPA DPRIVE program, Intel plans to design a dedicated integrated circuit accelerator (ASIC) to reduce performance over current that is currently associated with full homomorphic encryption. When fully implemented, the accelerator could achieve a significant improvement in the performance of the FHE workload over the CPU-driven systems, which may be reduces cryptogram processing time by five orders of magnitude.
With its experience in cloud infrastructure, software stacks and fully homomorphic encryption, Microsoft will be a critical partner in accelerating the commercialization of this technology when completed, enabling sharing free data and collaboration while encouraging privacy throughout the data lifecycle.
“We are delighted to bring our expertise in cloud computing and homomorphic encryption to the DARPA DPRIVE program, working with Intel to advance this transformative technology when ready for commercial applications that will help closing our customers’ last gap in data confidentiality —– keeping data completely secure and private, whether in storage, transfer or use, ”said Dr William Chappell, chief technology officer, Azure Global, and vice president, Mission Systems, Microsoft.
What’s next: The multipurpose DARPA DPRIVE program spans several stages starting with the design, development and validation of embedded IP blocks that will be integrated into a system-on-chip and full software stack. Throughout the project, Intel evaluates progress against pre-established performance targets based on artificial intelligence training and decision-making workloads using homomorphically encrypted data at scale.1 Beyond developing the key technologies required for accelerator design, Intel and Microsoft will work with international standards bodies to develop international standards for FHE. Intel will continue to invest in ongoing academic research in the field.