Cloud Storage Partitioning for 3D Data Retrieval
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for storing and retrieving 3-dimensionally organized data, such as seismic exploration data, face challenges in efficiency due to large data sizes and proprietary nature, leading to slow access times and difficulties in collaboration and data transfer across locations.
Innovation Solution
The method involves partitioning 3-dimensional data sets into smaller groups and storing them across multiple storage nodes in a cloud storage array, allowing for parallel reading and writing, and using a caching system to quickly retrieve data images by determining availability in local and remote caches.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If data is stored on disc in logical order based on x-axis, then reading data along x-axis is quick and efficient, but reading data along y-axis or z-axis requires time-consuming disc seeks
Solution Approach 1:
The patent segments the 3D data set into multiple independent data partitions or bricks, each stored separately in the cloud storage array. This segmentation allows different partitions to be accessed independently and in parallel, eliminating the sequential disc seek operations required when viewing data along different axes. Each partition can be retrieved simultaneously through distributed access points in the cloud architecture.
Solution Approach 2:
The patent transitions from traditional 2D disc-based storage to a 3D cloud storage architecture, adding a temporal/distributed dimension to data access. Data is organized in a 3D structure (x, y, z coordinates) and stored across multiple cloud nodes, enabling parallel access along any axis simultaneously. This dimensional transformation allows the system to overcome the fundamental limitation of sequential disc seeking by providing multiple concurrent access paths.
2Speed
If massive 3D data sets are stored locally, then fast access is achieved, but data security risks increase and collaboration across locations becomes difficult
Solution Approach 1:
The patent segments the complete 3D data set into multiple partitions distributed across different cloud storage nodes and geographic locations. This segmentation inherently improves security because no single location contains the complete sensitive data set. Even if one node is compromised, the full data remains secure. The segmentation also enables controlled access where different users can be granted permissions to specific partitions based on their needs.
Solution Approach 2:
The patent introduces a cloud storage intermediary layer that sits between the user and the raw data storage. This cloud infrastructure acts as a secure mediator, implementing authentication, authorization, encryption, and access control policies. The intermediary manages data retrieval requests, ensures proper security protocols are followed, and coordinates parallel access across multiple nodes without exposing the underlying data to unauthorized access.
3Reliability
If proprietary data is handled with care and not duplicated, then data security is maintained, but data transfer and collaboration are hindered
Solution Approach 1:
The patent creates controlled copies of data partitions across multiple cloud storage nodes and geographic locations. These copies are not unauthorized duplications but rather distributed replicas managed by the cloud system, enabling simultaneous access by multiple users from different locations. The copying mechanism maintains data integrity through checksums and version control while allowing collaborative work. Users can access identical or different partitions in parallel without compromising the original data's security or proprietary status.
Data Source
AI summary
A cloud based storage system and methods for uploading and accessing 3-D data partitioned across distributed storage nodes of the system. The data cube is processed to identify discrete partitions thereof, which partitions may be organized according to the x (e.g., inline), y (e.g., crossline) and/or z (e.g., time) aspects of the cube. The partitions are stored in unique storage nodes associated with unique keys. Sub-keys may also be used as indexes to specific data values or collections of values (e.g., traces) within a partition. Upon receiving a request, the proper partitions and values within the partitions are accessed, and the response may be passed to a renderer that converts the values into an image displayable at a client device. The request may also facilitate data or image access at a local cache, a remote cache, or the storage partitions using location, data, retrieval, and/or rendering parameters.