Data Segregation via Shard Distribution and Key Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
As network computing and data storage systems become more distributed, controlling access to data across multiple geographic regions becomes increasingly complex, posing challenges in preventing unauthorized access and ensuring data sovereignty.
Innovation Solution
Implementing redundancy encoding techniques, such as bundle-encoded and erasure-encoded data distribution, along with geographic segregation of shards and cryptographic key management, to ensure that only authorized entities can access the original data, while prohibiting unauthorized access by limiting the availability of shards and cryptographic information across regions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If data is distributed across multiple geographic regions, then data availability and accessibility are improved, but controlling access and preventing unauthorized reconstruction becomes more complex
Solution Approach 1:
The patent divides data into multiple shards and distributes them across different geographic regions. Each shard is a fragment of the original data, and no single shard contains the complete information. This segmentation allows data to be stored in multiple locations for improved availability while maintaining security, as unauthorized parties would need to collect all shards to reconstruct the original data.
Solution Approach 2:
The patent introduces cryptographic keys as an intermediary mechanism to control access to shards. The key management service acts as a mediator that distributes cryptographic keys to authorized entities. This intermediary layer simplifies access control by managing permissions centrally, allowing the system to maintain both data availability across regions and controlled access through a single point of key management.
2Ease of operation
If cryptographic information is distributed across multiple regions, then data accessibility is improved, but security against unauthorized access deteriorates
Solution Approach 1:
The patent segments cryptographic information into multiple parts and distributes them across different regions alongside the corresponding data shards. This ensures that cryptographic information is as distributed as the data itself, maintaining accessibility while preserving security through the same fragmentation mechanism that protects the data.
Solution Approach 2:
The patent implements local quality by associating specific cryptographic keys with specific geographic regions. Each region has the cryptographic information needed to access its local shards, but not the complete set of keys required to access all data. This localized distribution of cryptographic information enables region-specific access control while maintaining overall system security.
3Productivity
If all shards are made available in every region, then data reconstruction capability is improved, but unauthorized data access becomes more likely
Solution Approach 1:
The patent segments data into shards and strategically distributes them across geographic regions such that not all shards are available in every region. This segmentation ensures that while data can be reconstructed when needed (by gathering shards from multiple regions), unauthorized parties cannot easily access all shards in a single location to perform unauthorized reconstruction.
Solution Approach 2:
The patent implements partial action by providing only the necessary subset of shards to each region, rather than distributing all shards everywhere. Each region receives enough shards to contribute to data reconstruction when needed, but not enough to independently reconstruct the entire dataset. This partial distribution maintains productivity while reducing unauthorized access risk.
Data Source
AI summary
A computer system, such as a data storage system, implements techniques for segregating and controlling access to data stored in multiple regions. In some embodiments, redundancy coded shards generated from the data and stored in durable storage of a data storage system is allocated across multiple regions, but in a fashion that prevents actors with access to regions outside that of a “home” region from recovering a sufficient number of unique shards to regenerate the data represented thereby. In some embodiments, encryption is used to segregate the data by encrypting the generated shards, then storing the cryptographic information on or otherwise controlling access on hosts or other devices of only the home region.