Modular Storage Architecture for Scalable Compute and Data Access
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
As data storage systems scale, communication delays and computing power strain become significant issues due to increased components and memory, leading to bottlenecks and longer data access times, especially when expanding from a single chassis to a multi-chassis system.
Innovation Solution
The implementation of a storage system architecture that allows for independent scaling of compute resources and storage resources, utilizing disaggregated compute and storage resources with direct network-connected communication through switches, and employing non-volatile random access memory (NVRAM) as a buffer to improve latency and offload device management from storage drives.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If more components are added to increase storage capacity, then storage capacity is improved, but communication delays worsen due to fixed communication bandwidth
Solution Approach 1:
The system segments compute resources and storage resources into separate modular units that can be independently scaled. Compute modules and storage modules are physically separated and connected through standardized interfaces, allowing storage capacity to be increased without proportionally increasing compute resources or communication overhead in the storage expansion path.
Solution Approach 2:
A fabric module acts as an intermediary component that manages communications between compute modules and storage modules. This fabric module absorbs and manages communication bandwidth requirements, preventing communication bottlenecks when storage capacity is expanded by adding more storage modules.
2Quantity of substance
If more memory is added to increase storage capacity, then storage capacity is improved, but computing power becomes strained
Solution Approach 1:
The system divides resources into separate compute modules and storage modules. Storage modules contain their own local memory and caching resources, eliminating the need for compute modules to manage all memory operations. This segmentation allows storage capacity to be increased by adding storage modules without straining the computing power of individual compute modules.
Solution Approach 2:
Storage modules are self-sufficient units that manage their own memory, caching, and data operations independently. Each storage module has embedded controllers and local memory resources, allowing them to service storage requests without requiring proportional increases in external computing power.
3Quantity of substance
If storage system is expanded from single chassis to multi-chassis, then storage capacity is improved, but communication bottlenecks worsen
Solution Approach 1:
The system employs universal, standardized interfaces and protocols for communication between modules and across chassis boundaries. This universality allows storage modules in different chassis to communicate efficiently through the fabric module without protocol translation overhead, maintaining data access speed when expanding from single to multi-chassis configurations.
4Quantity of substance
If more components are added to increase capacity, then storage capacity is improved, but data access times increase
Solution Approach 1:
The system implements a nested hierarchy of storage resources with storage modules containing local memory, which in turn contains embedded controllers and caching layers. This nesting allows data to be accessed from multiple levels of the hierarchy, with frequently accessed data residing in faster local memory within storage modules, maintaining quick access times even as overall storage capacity increases through additional modules.
Data Source
AI summary
A storage system, blades, removable modules, and method of configuring a storage system are described. The storage system has blades with computing resources and storage resources. At least one of the blades has, or has added, one or more removable modules.