Distributed Smart Card Storage for Expansion and Recovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Local disks in elastic compute service instances have limited storage space, leading to low utilization and inability to expand laterally, and suffer from slow recovery in case of failures.
Innovation Solution
A distributed storage space management method that partitions and simulates storage space using hardware smart cards deployed on local and remote hosts, creating a unified namespace resource pool to manage storage resources flexibly and enable rapid recovery.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If local disk is used for storage, then storage access performance is improved (low latency, high IOPS), but storage space cannot be expanded laterally
Solution Approach 1:
The patent segments storage resources into local storage (for performance-critical data) and remote storage (for expandable capacity). The storage system is divided into multiple storage nodes that can be independently managed, allowing the system to scale by adding more remote storage nodes without affecting local storage performance.
Solution Approach 2:
The patent introduces a storage management intermediary layer that sits between the local disk and remote storage. This intermediary handles data distribution, caching, and access routing, enabling the system to maintain fast access speeds while providing access to expanded remote storage capacity.
2Speed
If local disk is used for storage, then storage performance is improved, but space utilization is low
Solution Approach 1:
The patent merges local storage and remote storage into a unified storage system. Data can be distributed across both local and remote storage resources, allowing the system to utilize both the high performance of local disks and the large capacity of remote storage, thereby improving overall space utilization.
Solution Approach 2:
The patent implements dynamic data placement and caching strategies where hot data is kept on local storage for fast access while cold data is moved to remote storage. The system dynamically adjusts data locations based on access patterns, optimizing both performance and space utilization.
3Speed
If local disk is used for storage, then storage access speed is improved, but recovery time in case of failure is slow
Solution Approach 1:
The patent implements redundant storage and data replication across multiple nodes before failures occur. When a local disk fails, the system can immediately retrieve data from remote storage or replicate data from other nodes, significantly reducing recovery time while maintaining fast access speeds during normal operation.
Data Source
AI summary
A distributed storage space management method, a computing device and a storage medium are provided. An embodiment is applied to a hardware smart card, which is deployed on a local host as a local hardware smart card. A corresponding storage space is partitioned for a local virtual device from a pre-configured storage device; the partitioned storage space is simulated to generate a simulated storage device corresponding to the local virtual device, for use by the local virtual device. A storage space application request of a remote hardware smart card is received, and according to the request, a corresponding storage space is partitioned for a remote virtual device from the pre-configured storage device, so that the partitioned storage space is simulated by means of the remote hardware smart card to generate a simulated storage device corresponding to the remote virtual device, for use by the remote virtual device.