Dynamic MDTS Allocation for Multi-Tenant SSD QoS

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing data storage devices, such as solid state drives (SSDs), face challenges in providing optimal quality of service (QoS) due to shared resources among multiple tenants, leading to suboptimal performance for both short and long commands, as they advertise a fixed maximum data transfer size (MDTS) that does not account for varying QoS requirements.

Innovation Solution

A data storage device with a controller that determines and allocates a maximum bandwidth to tenants, adjusting the MDTS and aggregated queue depth (QD) settings based on quality of service (QoS) requirements, allowing for differentiated MDTS per tenant and command type to optimize resource allocation and command processing.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If a fixed MDTS is advertised to all tenants, then device complexity is reduced and ease of operation is improved, but quality of service and resource allocation efficiency deteriorate

Engineering Contradiction:
ImproveQoS differentiationVSAvoidcontroller complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent implements dynamic MDTS adjustment where the controller modifies the maximum data transfer size parameter for each tenant based on real-time bandwidth allocation and QoS requirements. This allows the system to adapt resource allocation dynamically without requiring complex manual configuration, resolving the contradiction between adaptability and complexity.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The controller changes the MDTS parameter dynamically for different tenants based on their allocated bandwidth portions. By modifying this key parameter, the system achieves QoS differentiation and optimized resource allocation while maintaining a relatively simple controller architecture that only needs to adjust numerical values rather than restructure complex systems.

Inventive Principle:
Principle #35Parameter changes

2Productivity

If bandwidth is allocated to multiple tenants, then resource utilization is improved, but command processing performance deteriorates due to shared resources

Engineering Contradiction:
Improveresource utilizationVSAvoidcommand processing performance
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The patent segments the total interface bandwidth into distinct portions allocated to different tenants. The controller divides the maximum bandwidth by the number of tenants to determine individual bandwidth allocations, ensuring that each tenant receives a fair share while maintaining overall system productivity. This segmentation prevents any single tenant from monopolizing resources and degrading command processing performance.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The controller allocates bandwidth portions to tenants based on their specific needs and QoS requirements, providing partial bandwidth allocation rather than full allocation to each tenant. This approach allows multiple tenants to share resources effectively while ensuring that command processing performance is maintained through controlled resource distribution.

Inventive Principle:
Principle #16Partial or excessive action

3Productivity

If MDTS is increased for long commands, then data transfer efficiency is improved, but short command performance deteriorates due to resource contention

Engineering Contradiction:
Improvedata transfer efficiencyVSAvoidcommand execution efficiency
Core Design Contradiction:
ProductivityVSEase of operation

Solution Approach 1:

The patent applies different MDTS values to different tenants based on their specific QoS requirements and workload characteristics. Tenants with long data transfer workloads receive higher MDTS allocations for improved efficiency, while tenants with short commands receive appropriate MDTS settings that optimize their execution. This localized quality adjustment resolves the contradiction by tailoring parameters to specific usage patterns.

Inventive Principle:
Principle #3Local quality

Data Source

PatentUS11934684B2Maximum data transfer size per tenant and command type
Publication Date: 2024.03.19 SANDISK TECHNOLOGIES LLC
  • US11934684B2 patent drawing
  • US11934684B2 patent drawing
  • US11934684B2 patent drawing

AI summary

A data storage device includes a memory device and a controller coupled to the memory device. The controller is configured to determine a maximum bandwidth of an interface, allocate a portion of the maximum bandwidth to one or more tenants, either: determine a maximum data transfer size (MDTS) setting based on quality of service (QoS) requirements, determine an aggregated queue depth (QD) setting based on QoS requirements, or determine a combined MDTS and aggregated QD setting based on QoS requirements, and provide the determined settings to the one or more tenants.