Parallel Data Placement Buffer Sharing in Memory Sub-Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional memory sub-systems face increased write amplification and reduced performance due to inefficient garbage collection and lack of host control over data placement, leading to excessive use of fast memory resources for accumulation buffers.
Innovation Solution
Dynamically allocate accumulation buffers to data placement handlers on demand, utilizing backup spaces in storage media for time-sharing, and implementing a communication protocol that supports flexible direct placement (FDP) and zoned namespace (ZNS) to optimize resource usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a large number of data placement handlers are run in parallel to support host control of data placement, then host control capability and data placement flexibility are improved, but fast memory resource consumption for accumulation buffers increases excessively
Solution Approach 1:
Multiple data placement handlers share a single accumulation buffer through time-multiplexed allocation. The buffer is dynamically assigned to different handlers based on their current operational needs, allowing parallel handlers to cooperate and share fast memory resources rather than each requiring dedicated buffers.
Solution Approach 2:
The accumulation buffer allocation is dynamically adjusted based on real-time handler activity. When a handler needs the buffer, it is allocated; when not needed, it is deallocated and made available to other handlers. This dynamic time-sharing approach allows the system to support many more parallel handlers than the ratio of fast memory resources would otherwise permit.
2Device complexity
If conventional garbage collection is used without host control, then system simplicity is maintained, but write amplification increases and performance decreases
Solution Approach 1:
The host system provides feedback control over data placement by issuing directives to the memory sub-system about where data should be written. This feedback mechanism allows the host to optimize data placement patterns, enabling more efficient garbage collection and reducing write amplification while maintaining reasonable system complexity through standardized communication protocols.
Solution Approach 2:
A communication protocol acts as an intermediary between the host and memory sub-system, carrying data placement directives that guide the memory controller's garbage collection and data placement operations. This intermediary layer enables sophisticated control without requiring complex direct integration between host and memory management functions.
3Reliability
If dedicated accumulation buffers are allocated to each data placement handler, then handler operation reliability is improved, but resource allocation efficiency deteriorates
Solution Approach 1:
The accumulation buffer is periodically allocated to different data placement handlers in a time-multiplexed fashion. Each handler receives dedicated buffer access when it needs to perform accumulation operations, ensuring operational reliability, while the periodic reassignment to different handlers maximizes resource utilization efficiency.
Data Source
AI summary
A memory sub-system having a first memory configured as a non-volatile storage medium of the memory sub-system; and a second memory having an access speed faster than the first memory. A processing device of the memory sub-system is configured to: run a plurality of data placement handlers concurrently; allocate a plurality of data buffers from the second memory to the plurality of data placement handlers respectively; reserve a plurality of backup spaces from the first memory for the plurality of data placement handlers respectively; allocate accumulation buffers from the second memory; and arrange the data placement handlers to time share the accumulation buffers via usages of the backup spaces.