Entropy Distribution in Distributed Computing Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Distributed computing systems face entropy constraints when a large number of computing devices are added, as existing methods for generating entropy, such as hardware random number generators, become insufficient to manage the increased demand.
Innovation Solution
A method that utilizes a true random number generator in a management computer to generate entropy, which is then combined with entropy from a main computer and distributed to physical nodes via a hypervisor, ensuring efficient entropy management across a distributed computing system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If hardware random number generators are used to generate entropy in each computing device, then entropy can be generated locally, but the entropy supply becomes insufficient when a large number of computing devices are added to the distributed system
Solution Approach 1:
The patent combines entropy generation resources across multiple computing devices in a distributed system. Instead of each device relying on its own hardware random number generator, the system merges entropy generation capabilities centrally or across nodes, allowing the total entropy supply to scale with the number of devices in the system rather than being limited by individual device capabilities.
Solution Approach 2:
The patent creates a universal entropy supply mechanism that serves multiple computing devices simultaneously. A single entropy generation source or coordinated set of sources provides entropy to multiple devices, making the entropy generation capability multi-functional and scalable to any number of devices in the distributed system.
2Quantity of substance
If pseudo-random software random number generators are used, then entropy generation is sufficient for static systems, but the system lacks the capability to handle dynamic expansion
Solution Approach 1:
The patent transitions from static entropy generation (single device or fixed capacity) to dynamic entropy generation that adapts to system changes. The distributed entropy generation system can dynamically allocate and scale entropy resources as computing devices are added or removed from the system, maintaining sufficient entropy supply throughout the expansion process.
3Productivity
If more computing devices are added to increase system capacity, then processing power increases, but entropy becomes constrained
Solution Approach 1:
The patent segments the entropy generation function across multiple computing devices in the distributed system. Rather than having entropy generation bottlenecked at a single point, each device or group of devices can contribute to or draw from distributed entropy resources, allowing system processing capacity to scale without proportionally increasing entropy constraints.
Data Source
AI summary
In one embodiment, a method generates first entropy using a true random number generator in a management computer configured to manage a main computer in a computing device. The main computer controls a set of physical nodes including a set of services running in a set of virtual machines. The method then provides the first entropy to the main computer and the first entropy is combined with second entropy generated by the main computer to generate third entropy. The third entropy is provided to the set of physical nodes where the set of virtual machines access the third entropy via a hypervisor.