Object Storage Partitioning with Configurable Sequential or Hash Distribution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object storage services lack the ability to accommodate tenants' requirements for storing objects with different distribution characteristics, as they default to sequential distribution methods.
Innovation Solution
Implementing a customizable distribution method that allows tenants to select between sequential and discrete distribution manners, with discrete distribution using hash values and configurable rules to distribute objects across logical partitions based on name strings.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the object storage server stores objects in sequential distribution manner by default, then the storage structure is simple and easy to implement, but the system cannot meet tenants' requirements for storing objects with different distribution characteristics
Solution Approach 1:
The patent implements dynamic distribution by allowing the system to switch between sequential distribution and hash distribution modes based on tenant requirements. The distribution strategy is no longer fixed but can be dynamically selected through configuration parameters, enabling the system to adapt to different storage scenarios while maintaining operational simplicity when using default sequential distribution.
Solution Approach 2:
The patent introduces configurable parameters (distribution mode selection) that allow tenants to change the distribution behavior from sequential to hash-based distribution. By modifying the distribution parameter based on tenant needs, the system achieves versatility without permanently increasing structural complexity, as the core storage mechanism remains the same while the distribution strategy is adjustable.
2Productivity
If the object storage server uses a single sequential distribution manner, then the implementation is straightforward, but the data distribution across partitions may be unbalanced for certain object naming patterns
Solution Approach 1:
The patent allows switching between distribution modes via configuration parameters. When object naming patterns cause poor sequential distribution, the system can change the distribution parameter to hash-based distribution, which provides more uniform data spread across partitions. This parameter change improves storage efficiency without requiring a complete redesign of the storage architecture.
Solution Approach 2:
The system dynamically adapts its distribution strategy based on the chosen configuration. Rather than being locked into a single static distribution method, the system can dynamically select between sequential and hash distribution modes, optimizing storage efficiency for different workload characteristics while keeping the implementation manageable through configurable parameters.
Data Source
AI summary
The present disclosure relates to object storage service-based storage methods and apparatuses. In one example method, an object storage server pre-obtains a configuration parameter, where the configuration parameter includes a target distribution manner selected by a tenant from a plurality of distribution manners, and the plurality of distribution manners include a sequential distribution manner and a discrete distribution manner. The object storage server creates a bucket based on the configuration parameter, where the bucket includes a plurality of logical partitions. When the object storage server obtains an object to be stored in the bucket, the object storage server may select a target logical partition from the plurality of logical partitions based on the configuration parameter and a name string of the object to store the object in the target logical partition.