DISAGGREGATED DATA CENTER PLANNING USING HOMOMORPHIC ENCRYPTION

Smart overview of the Invention

The patent describes a method for configuring disaggregated data centers using homomorphic encryption. These data centers consist of separate compute nodes and distributed memory nodes, which communicate over a network. The technique involves processing homomorphically-encrypted logs of memory accesses. By analyzing these logs, the system can optimize the allocation of memory resources to compute nodes, enhancing the data center's efficiency and scalability while maintaining data privacy.

Background

Disaggregated data centers offer the advantage of flexibility and scalability by decoupling compute, storage, and networking resources. However, configuring these centers to handle diverse workloads is technically challenging. The invention addresses these challenges by providing a method that supports resource scalability, manages complexity, and preserves data privacy through advanced encryption techniques.

Technical Summary

The method involves receiving homomorphically-encrypted logs from distributed memory nodes via a communication network. These logs are processed to compute a function set, resulting in a result set that is partially distributed back to the memory nodes. The plaintext version of this result set is then requested from the nodes. Using this information, the system determines the optimal assignment of memory resources to compute nodes, thus configuring the data center efficiently.

Implementation Details

The patent provides a detailed description of the system's implementation, which includes narrative text, flowcharts, and block diagrams. The method can be executed on various computing environments, including desktops, laptops, tablets, and cloud-based systems. The system utilizes a combination of volatile and persistent storage media to execute the computer-readable program instructions that perform the inventive methods.

System Architecture

The computing environment comprises components such as processors, communication fabrics, volatile memory, and persistent storage. These components work together to execute the data center orchestrator, which manages the configuration process. The system is designed to work with various types of storage media and can operate in both public and private cloud environments, ensuring flexibility and adaptability to different technological setups.