Gateway Device Converts NVMe Packets to RDMA for Storage Performance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current NVMe over fabrics (NOF) protocol-based storage solutions suffer from low performance due to the inferior performance of hard disks and complex instruction sets, leading to inefficiencies in read/write operations.
Innovation Solution
A packet processing method that converts NVMe instructions into RDMA instructions, allowing for direct memory operations on RDMA storage nodes, which offers faster read/write speeds and simpler instruction sets, thereby improving storage performance and reducing operational complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If hard disk is used for storage in NOF protocol-based solution, then storage capacity is provided, but performance is inferior and instruction set is complex
Solution Approach 1:
The patent changes the storage medium parameter from hard disk to memory, and changes the protocol parameter from NVMe to RDMA. This parameter substitution transforms the storage system to use memory with simpler instruction sets and faster access speeds, directly resolving the contradiction between storage performance and instruction complexity
Solution Approach 2:
The gateway device acts as an intermediary that translates NVMe instructions into RDMA instructions. This mediator converts the complex NVMe protocol operations into simpler RDMA operations, allowing the client to continue using NVMe interface while the actual storage operations are performed using efficient RDMA commands on memory
2Speed
If NVMe instructions are used for hard disk operations, then storage access is enabled, but read/write speed is limited
Solution Approach 1:
The patent changes the storage medium parameter from hard disk to memory, which inherently provides faster read/write speeds. Simultaneously, it changes the operation parameter from NVMe instructions to RDMA instructions, which are optimized for high-speed memory access, thereby achieving both speed improvement and operation simplification
3Productivity
If NOF protocol is used for network storage access, then remote storage is enabled, but CPU and network pressure increases
Solution Approach 1:
The gateway device serves as an intermediary that handles protocol translation and instruction conversion. By offloading the complex NVMe to RDMA translation task to the gateway, the client CPU is relieved from processing complex storage protocols, reducing CPU pressure and improving storage access efficiency
Solution Approach 2:
The patent replaces the NVMe protocol mechanism with the RDMA protocol mechanism. RDMA enables direct memory access that bypasses the CPU, substituting the CPU-intensive NVMe processing with a more efficient memory-direct access mechanism, thereby reducing CPU and network pressure
Data Source
AI summary
The technology of this application relates to a packet processing method, a gateway device, and a storage system, and belongs to the field of storage technologies. In this disclosure, a request for accessing a non-volatile memory express (NVMe) node is converted into a request for accessing a remote direct memory access (RDMA) node. Because a storage medium of the NVMe node is a hard disk, and a storage medium of the RDMA node is a memory, the memory can provide a faster read/write speed than the hard disk. Therefore, storage performance is improved. In addition, this helps a conventional NVMe over fabrics (NOF) storage system expand an RDMA memory pool, and improves flexibility of networking and capacity expansion of the storage system.