Virtualizing Key Hierarchy Using Partially-Oblivious Pseudorandom Function
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing key management systems face limitations in scalability and security, particularly when dealing with large numbers of encryption keys, as they are often restricted by memory capacity in Hardware Security Modules (HSMs) and require trust in third-party service providers, making them vulnerable to malicious behavior.
Innovation Solution
The implementation of Partially-Oblivious Pseudorandom Functions (P-OPRFs) allows for the virtualization of keys, enabling a single key to be used to generate multiple distinct keys for different users, ensuring secure key management without revealing the original key to service providers, thus enhancing scalability and security.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a Hardware Security Module (HSM) is used to store encryption keys, then key security is improved, but the number of keys that can be stored is limited due to memory constraints
Solution Approach 1:
The key management system is segmented into multiple HSM instances, each storing a portion of the key hierarchy. This allows the system to manage large numbers of keys by distributing them across multiple secure storage units, overcoming the memory limitations of individual HSMs while maintaining security through cryptographic segmentation of the key space.
Solution Approach 2:
The system transitions from storing keys in a single-dimensional memory structure to a multi-dimensional key hierarchy where keys are organized across multiple levels and HSM instances. This dimensional expansion allows exponential growth in key capacity without linearly increasing memory requirements in any single HSM.
2Adaptability or versatility
If more encryption keys are stored to service more users, then system versatility is improved, but memory requirements increase and scalability becomes difficult
Solution Approach 1:
A single HSM instance is designed to perform multiple functions by servicing different key hierarchies and user groups simultaneously. The HSM can derive and manage keys for multiple tenants or applications using a single root key, eliminating the need for separate HSM instances for each key hierarchy and reducing overall memory requirements.
Solution Approach 2:
The system changes the parameter of key derivation by using cryptographic functions that can generate unlimited unique keys from a single root key. This allows the system to service any desired number of users with unique keys while maintaining constant memory footprint in the HSM, as keys are derived on-demand rather than stored.
3Ease of operation
If a single root key is used to service multiple users, then key management simplicity is improved, but security is reduced if the service provider is breached
Solution Approach 1:
The key hierarchy is segmented into customer-controlled root keys and service-provider-managed derived keys. This segmentation allows simple key management from the customer perspective while maintaining security, as the service provider only handles derived keys that cannot reconstruct the customer's root key, protecting against provider breaches.
Solution Approach 2:
A cryptographic intermediary function (key derivation function) is introduced between the root key and user-specific keys. This intermediary allows the service provider to manage user keys without ever accessing the root key, creating a security barrier that protects against provider breaches while maintaining operational simplicity.
4Ease of manufacture
If traditional key storage methods are used, then implementation is straightforward, but post-quantum security is compromised
Solution Approach 1:
The system replaces traditional physical key storage mechanisms with quantum-resistant cryptographic primitives. By using lattice-based or post-quantum key derivation functions instead of classical cryptographic methods, the system maintains implementation simplicity while achieving post-quantum security, as the mathematical hardness assumptions remain valid against quantum computers.
Data Source
AI summary
A computing device includes an interface configured to interface and communicate with a communication system, a memory that stores operational instructions, and processing circuitry operably coupled to the interface and to the memory that is configured to execute the operational instructions to perform various operations. The computing device processes an input value (e.g., associated with a key) based on a blinding key (e.g., homomorphic encryption) to generate a blinded value and generates an Oblivious Key Access Request (OKAR). The computing device transmits the OKAR to another computing device (e.g., associated with a Key Management System (KMS) service) and receives a blinded key therefrom that is based on a Partially-Oblivious Pseudorandom Function (P-OPRF). The computing device processes the blinded key based on the blinding key (e.g., homomorphic decryption) to generate the key (e.g., associated with the input value). In some examples, the computing device accesses secure information based on the key.