Hash Table Remote Memory Operations Without Host Intervention
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network communication methods for performing complex operations in remote memory incur high latency and require remote host involvement, making them inefficient and suboptimal for distributed applications.
Innovation Solution
Implementing Remote Direct Memory Operations (RDMO) commands, such as Maximum Compare-and-Swap (MAX-CAS), Hash Table get/set, Table Append, and fault-tolerant remote logging, directly executed in remote memory by network devices without host intervention, using RDMA or similar transport protocols.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing network communication methods are used to perform complex operations in remote memory, then remote host involvement is required, but latency increases and efficiency decreases
Solution Approach 1:
The patent extracts the host involvement from the remote memory operation process. Network devices are empowered to execute complex operations (compare-and-swap, hash table operations, logging) directly in remote memory without requiring remote host intervention. This extraction eliminates the latency associated with host processing while maintaining atomic execution guarantees through dedicated network device implementation of these operations.
Solution Approach 2:
The patent implements self-service by enabling network devices to autonomously perform complex memory operations. Instead of requiring remote hosts to process and manage these operations, the network devices themselves execute compare-and-swap, hash table get/set, append, and logging operations directly in the remote memory, reducing latency and improving efficiency.
2Reliability
If remote host involvement is used for complex memory operations, then operation completeness is ensured, but communication overhead increases
Solution Approach 1:
The patent extracts the operation execution from the remote host to the network device. By enabling network devices to directly execute complex operations in remote memory, the system eliminates unnecessary communication rounds between hosts and network devices, reducing communication overhead while ensuring operation completeness through atomic execution at the network device level.
Solution Approach 2:
The patent introduces network devices as intermediaries that can autonomously execute complex operations. These intermediaries (network devices) receive operations from hosts and complete them directly in remote memory without requiring further host involvement, thereby reducing communication overhead while maintaining operation reliability.
3Productivity
If simple RDMA operations are used, then network device execution is efficient, but complex operations cannot be performed
Solution Approach 1:
The patent extends the functionality of network devices to handle multiple types of operations efficiently. By implementing compare-and-swap, hash table operations (get, set, append), and logging operations directly in network devices, the system achieves multi-functionality without sacrificing execution efficiency. Network devices become universal processors capable of handling both simple and complex operations with the same efficiency as traditional RDMA operations.
Data Source
AI summary
A system includes a first network device and a second network device. The first network device is to send over a network a command that (i) specifies a key for accessing a hash table in a memory and (ii) instructs that a value be read or written at a location in the hash table corresponding to the key. The second network device is to receive the command over the network, and to execute the command by calculating the location in the hash table based on the key, and reading or writing the value at the location.