Confidential Computing using AMD SEV SNP Hardware Mediated Enclaves

This CPU level enforced feature encrypts all of the contents of the virtual machine in memory and over-the-bus, thus providing the virtual machine with an unparalleled level of confidentiality and integrity. This protects virtual payloads from the hypervisor as well as side-channel attacks that might stem from malicious co-tenants within a server instance.

AES Encryption

The Array Insights platform is a true end-to-end encryption platform, where everything is encrypted. This includes:1. Data at Rest (i.e. datasets, database records, virtual encrypted disks within virtual machines).2. Data in Motion (i.e. all network connections use TLS 1.2 with ECC 384 PFS).3. Data in Use (i.e. all virtual machines are hosted within Azure Confidential Compute instances). All datasets and database are encrypted using AES 256 GCM, thus ensuring the confidentiality and integrity of the data. All TLS 1.2 connections use AES 256 for encryption and ECC 384 PFS for endpoint verification and key exchange. AMD SEV SNP and virtual encrypted disks within virtual machines use AES 128 for performance purposes.

Digital Signatures

The Array Insights Platform makes extensive use of digital signatures in order to ensure the identity and integrity of data throughout the platform. All digital signatures use ECC 384, although the platform is designed to allow for configurable cipher-suites. Digital signatures are used for:1. Datasets that encrypted and digitally signed.2. Remote attestation certificates that are digitally signed.3. All TLS 1.2 connections. TLS 1.2 using ECC 384 PFS with frequently rotated digital certificates All network connections between distributed endpoints within the Secure AI Labs Platform are encrypted using TLS 1.2, using ECC 384 PFS. Each virtual machine instance (i.e. Secure Computation VM, API Portal Instance, Database Portal Instance) generate their own short lived and ephemeral digital certificates. Digital certificates for TLS 1.2 are rotated frequently (i.e. every 20 minutes)

Greater breadth powers more depth

With our Data Federation Platform, discover results you can be confident in. This is especially critical for rare conditions, minority populations, and personalized medicine breakthroughs.