Disaggregated Storage Network Elastic Scaling

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Traditional storage networks are inefficient in utilizing less expensive hardware solutions due to architectural limitations, leading to high computational overhead and power consumption, as they require complex hardware with multiple CPU cores and dedicated RAM, and are not scalable to meet varying data processing demands.

Innovation Solution

The implementation of a disaggregated storage network architecture that decouples storage and data services, using low-power, low-cost Fabric Attached Bunch of Flash (FBOF) devices with programmable System on a Chip (SoC) solutions, and data service nodes with high processing capabilities, allowing flexible and scalable storage targets and dynamic reconfiguration of resources.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If traditional storage networks use complex hardware with multiple CPU cores and dedicated RAM, then processing capability is improved, but cost and power consumption increase

Engineering Contradiction:

Improveprocessing capabilityVSAvoidpower consumption

Core Design Contradiction:

ProductivityVSUse of energy by stationary object

Solution Approach 1:

The patent segments the storage network into separate functional components: storage targets (FBOF devices) that handle data storage and data service nodes that handle processing tasks. This segmentation allows each component to be optimized independently, enabling the use of low-power FBOF devices for storage while distributing processing requirements across multiple service nodes, thereby reducing overall power consumption while maintaining processing capability.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent extracts the processing functions from the storage targets and places them in separate data service nodes. This extraction allows the storage targets to use simple, low-power FBOF hardware without the need for expensive multi-core CPUs and dedicated RAM, while the processing requirements are fulfilled by the data service nodes that can be dynamically scaled and configured.

Inventive Principle:

Principle #2Taking out (Extraction)

2Productivity

If traditional storage networks use complex hardware with multiple CPU cores and dedicated RAM, then processing capability is improved, but cost increases

Engineering Contradiction:

Improveprocessing capabilityVSAvoidcost

Core Design Contradiction:

ProductivityVSEase of manufacture

Solution Approach 1:

By segmenting the storage network into storage targets and data service nodes, the patent enables the use of inexpensive FBOF devices for storage while handling processing requirements through software-based data service nodes. This approach eliminates the need for expensive specialized storage hardware with multiple CPUs and dedicated RAM, significantly reducing the cost of deployment while maintaining adequate processing capability through virtualized service nodes.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent employs inexpensive FBOF devices as storage targets that can be easily deployed and replaced. These devices lack the expensive processing components of traditional storage systems but are sufficient for their storage function, with processing capabilities provided by shared data service nodes. This approach prioritizes cost-effectiveness while meeting processing requirements through resource sharing.

Inventive Principle:

Principle #27Cheap short-living objects (Disposable)

3Reliability

If traditional storage networks are configured for worst-case scenarios, then reliability is improved, but computational resource waste increases

Engineering Contradiction:

Improvesystem reliabilityVSAvoidcomputational resource waste

Core Design Contradiction:

ReliabilityVSLoss of energy

Solution Approach 1:

The patent implements dynamic resource allocation where data service nodes can be dynamically created, scaled, and terminated based on actual workload demands. This dynamic approach replaces the static worst-case provisioning of traditional systems, allowing the network to maintain reliability by allocating processing resources only when needed, thereby eliminating the computational resource waste associated with always-running under-provisioned hardware for peak scenarios.

Inventive Principle:

Principle #15Dynamics

4Device complexity

If storage and processing are combined in the same hardware, then device simplicity is improved, but scalability and flexibility deteriorate

Engineering Contradiction:

Improvedevice simplicityVSAvoidscalability

Core Design Contradiction:

Device complexityVSAdaptability or versatility

Solution Approach 1:

The patent segments storage and processing into separate functional units: simple FBOF storage targets and software-based data service nodes. This segmentation maintains device simplicity at the storage target level while enabling scalability through the independent provisioning and configuration of data service nodes. The separation allows storage capacity to be scaled by adding FBOF devices and processing capacity to be scaled by adding or configuring service nodes, providing flexibility without increasing individual device complexity.

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS11405455B2Elastic scaling in a storage network environment

Publication Date: 2022.08.02 OVH US LLC

AI summary

Disaggregated storage clusters are disclosed. These disaggregated storage clusters include a plurality of storage targets coupled to each other through a switch and including storage targets including storage and data services storage targets. Data and requests can for storage areas maintained by the storage cluster can be routed between the target of the storage clusters based on pipeline definitions for those storage areas.