Switched Direct Attached Storage Architecture Without RDMA
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional data storage architectures, such as Clustered DAS and NAS/SAN, fail to efficiently address the growing needs of data storage and analysis by lacking scalability, accessibility, and performance while increasing costs, and they introduce unnecessary protocol overhead.
Innovation Solution
The proposed Switched Direct Attached Shared Data Storage Architecture leverages packet switching, centralized physical storage media, and distributed storage processing software to enable scalable, accessible, and high-performance data storage by using NVMe, SAS/SATA, or SCSI over PCIe protocols, and encapsulates IO operations into standard Ethernet frames, eliminating redundant protocol conversions and enhancing data accessibility and performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional data storage architectures (Clustered DAS or NAS/SAN) are used to enable multiple hosts to access storage devices, then data accessibility is improved, but protocol overhead increases and performance deteriorates
Solution Approach 1:
The patent extracts and removes the RDMA protocol layer from the storage access architecture. By eliminating RDMA and using standard Ethernet frames with TCP/IP protocols instead, the system removes the complexity of specialized hardware and protocols while maintaining multi-host access capability. The storage devices are accessed through conventional network protocols rather than requiring dedicated RDMA infrastructure.
Solution Approach 2:
The patent makes the storage access system universal by using standard Ethernet and TCP/IP protocols that can be used by any host with standard network interfaces. This eliminates the need for specialized RDMA hardware and software on each host, allowing any standard networked computer to access the storage devices without requiring proprietary components.
2Adaptability or versatility
If conventional data storage architectures are used to provide shared storage access, then data accessibility is improved, but costs increase
Solution Approach 1:
The patent replaces expensive specialized RDMA hardware and software components with inexpensive standard Ethernet network interfaces and conventional TCP/IP software stacks. Instead of requiring costly RDMA-capable network cards and specialized drivers on each host, the system uses affordable standard network infrastructure that is already widely deployed in enterprise environments.
3Quantity of substance
If scalable storage architectures are implemented to meet growing data storage needs, then storage capacity is improved, but protocol overhead increases and performance deteriorates
Solution Approach 1:
The patent substitutes the mechanical/specialized RDMA protocol mechanism with standard Ethernet TCP/IP networking mechanisms. Instead of using the specialized RDMA protocol path that requires specific hardware acceleration and complex protocol handling, the system uses conventional network protocols that can be processed by standard network stacks, thereby maintaining performance while enabling scalability.
Data Source
AI summary
A method and apparatus for accessing multiple storage devices from multiple hosts without use of remote direct memory access (RDMA) as disclosed herein include: providing a data store switch fabric enabling data communications between a data storage access system and a plurality of compute nodes, each compute node having integrated compute capabilities, data storage, and a network interface controller (Host NIC); providing a plurality of physical data storage devices; providing a host bus adapter (HBA) in data communication with the plurality of physical data storage devices and the plurality of compute nodes via the data store switch fabric, the HBA including at least one submission queue and a corresponding shadow queue; receiving an input/output (I/O) request from the plurality of compute nodes; including an element of the I/O request to the at least one submission queue; and including additional information related to the element of the at least one submission queue to the corresponding shadow queue.