Network Card Throughput Optimization via Payload Data Skipping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current network testing systems face bottlenecks in throughput due to time-consuming data unit payload copying and processing overhead when transmitting and receiving large data units, particularly in network communication optimization.
Innovation Solution
Implementing a software interface with flags like READIGNORE and WRITEIGNORE to eliminate payload data transfer between memory and network communications units, allowing dummy data or no data to be used, and reusing payload data across multiple units, thereby reducing processing overhead and increasing throughput.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If payload data is copied between memory and network communications unit for each data unit, then data integrity and protocol compliance are ensured, but throughput decreases due to time-consuming copying operations
Solution Approach 1:
The patent extracts the payload copying operation from the mandatory processing sequence by introducing a flag mechanism that allows the payload to be skipped entirely when using dummy data. This removes the harmful copying action while preserving necessary header processing and protocol compliance checks.
Solution Approach 2:
The patent changes the state of payload data from requiring full copying to allowing selective copying by introducing a parameter (flag) that controls this behavior. When the flag is set, the system transitions from copying actual payload data to using dummy data or skipping payload processing entirely.
2Reliability
If complete payload data is transferred for each data unit, then network protocol requirements are met, but processing overhead increases due to redundant operations
Solution Approach 1:
The patent applies partial action by transferring only the necessary header information and metadata while omitting the payload portion when dummy data is used. This reduces processing overhead while maintaining protocol compliance for testing scenarios where payload content is not critical.
Solution Approach 2:
The patent uses copying of header and metadata information without copying the payload data when the flag is set. This selective copying approach maintains protocol structure requirements while avoiding redundant payload transfer operations.
3Measurement precision
If payload data is processed and transferred for every data unit, then data communication accuracy is maintained, but system throughput is limited by memory transfer bandwidth
Solution Approach 1:
The patent extracts the payload from the data transfer process when dummy data is indicated, separating it from the essential header and metadata transfer. This eliminates the memory transfer bottleneck for large payloads while maintaining accuracy for protocol-critical fields.
Solution Approach 2:
The patent performs partial data transfer by sending only the critical header and metadata portions of data units when payload copying is disabled. This maintains communication accuracy for protocol validation while dramatically reducing memory transfer requirements and increasing throughput.
Data Source
AI summary
Data communication optimization is disclosed. In a method, when data transmission is requested, an outgoing data unit is prepared with dummy payload data. In a method, when data reception is requested, the header and validation information from an incoming data unit is provided while the payload of the data unit is ignored. The methods cause an increase in the throughput of a network card when transmitting and/or receiving a data units having a large payload. The methods may be achieved on a hardware device included on a board in a network testing system and/or with associated software.