Transport Offload Engine Metadata for Network I/O Bottlenecks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing communication speed in systems creates an input/output bottleneck as processors, designed primarily for computing, struggle to keep up with data flow, leading to inefficient data processing and high memory requirements for transport offload engines, making cost-effective data transmission challenging.
Innovation Solution
A data object is used to facilitate network communication management between network protocol layers, allowing the transport offload engine to process data associated with upper protocol layers, reducing the need for large memory and enabling efficient data transmission by utilizing metadata, indicators for byte and CRC calculations, and markers for PDU alignment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If the processor handles data transmission directly, then processing capability is sufficient, but the I/O speed cannot keep up with network data flow
Solution Approach 1:
The patent extracts the I/O processing function from the processor by introducing a transport offload engine. The processor generates data lists and transmits them to the TOE, which then handles data retrieval from memory and network transmission. This separation allows the processor to focus on computing tasks while the TOE handles high-speed I/O operations, resolving the contradiction between transmission speed and processing efficiency.
2Productivity
If large memory is provided for data lists and control blocks in the TOE, then data transmission capability is improved, but the cost and silicon die area increase significantly
Solution Approach 1:
The patent uses metadata objects that contain references or descriptors pointing to data in host memory rather than storing actual data copies in TOE memory. The metadata includes information such as data location, length, and control information, which allows the TOE to access and process data without requiring large onboard memory capacity. This copying approach maintains data transmission capability while significantly reducing the memory size required in the TOE.
Data Source
AI summary
A system, method, and related data structure are provided for transmitting data in a network. Included is a data object (i.e. metadata) for communicating between a first network protocol layer and a second network protocol layer. In use, the data object facilitates network communication management utilizing a transport offload engine.