SCM Internal Volumes for Storage Cluster Memory Relief
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Nodes in a storage cluster face challenges in allocating sufficient memory resources to store performance monitoring data and metadata due to limited memory availability, leading to decreased monitoring objects and indicators, which cannot meet user requirements.
Innovation Solution
Create an internal volume using storage class memory (SCM) disks to store target data, allowing nodes to access and persistently store data efficiently, reducing the need for memory allocation and minimizing data loss.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If nodes allocate large memory to store performance monitoring data and metadata, then data storage capacity is improved, but memory resource availability deteriorates
Solution Approach 1:
The patent transitions from using volatile memory (RAM) to using persistent storage (SCM disks) for storing performance monitoring data and metadata. This dimensional change in storage medium allows nodes to access data from a shared storage system instead of relying on limited local memory, thereby expanding storage capacity without consuming node memory resources.
Solution Approach 2:
The shared storage system serves multiple functions: it stores performance monitoring data, metadata, and other large volume data for multiple nodes simultaneously. This universal storage resource replaces the need for each node to maintain separate memory allocations, improving overall storage capacity while reducing individual node memory requirements.
2Speed
If nodes use local memory for data storage, then data access speed is improved, but storage scalability deteriorates
Solution Approach 1:
The patent introduces a shared storage system with caching mechanisms as an intermediary between nodes and persistent storage. The caching layer provides fast access to frequently accessed data while the backend storage system provides scalable capacity. This intermediary structure maintains data access speed comparable to local memory while enabling unlimited storage scalability.
3Measurement precision
If nodes allocate more memory for metadata storage, then monitoring precision is improved, but memory availability deteriorates
Solution Approach 1:
The patent extracts metadata storage from node memory to the shared storage system. By taking out metadata from local memory and storing it externally, nodes can maintain detailed and precise monitoring information without the metadata consuming valuable local memory resources. This extraction enables high monitoring precision while preserving memory availability for other critical functions.
Data Source
AI summary
The embodiments of the present application disclose a data storage method and apparatus, a device, a medium and a storage cluster, which relate to the field of storage technologies and can meet a data storage requirement for nodes in the storage cluster. The method includes the following steps: accessing an internal volume of the storage cluster, and writing its own target data into the internal volume for storage, wherein the internal volume is created through the following step: creating the internal volume on the basis of a plurality of storage class memory disks configured for the storage cluster, wherein the internal volume is configured to store respective target data of each node in the storage cluster.