Jumbo Frame Translation via Network Buffer Pool Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Implementing Jumbo frames over existing network stacks is challenging due to differences in Ethernet frame sizes across networks, leading to inefficiencies in data transmission and increased overhead, as applications may not be aware of the specific network's frame sizes and require modifications to handle varying packet sizes.
Innovation Solution
The system introduces a new pool of network buffers with larger sizes to support Jumbo and non-Jumbo frames, using a scatter-gather mechanism to form and transmit packets, allowing seamless communication between networks with different MTU sizes without modifying the network stack. This involves partitioning data into segments, storing them in corresponding buffers, and reassembling packets at the TCP layer to maintain compatibility with applications unaware of packet chaining.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If standard Ethernet frame sizes are used, then compatibility with existing network stacks is maintained, but network efficiency deteriorates due to increased overhead and more packets required for large data transmission
Solution Approach 1:
The patent introduces a Jumbo Frame Translation (JFT) device as an intermediary between networks with different MTU sizes. This device receives jumbo frames from one network, segments them into standard-sized Ethernet frames, and forwards them to networks that cannot handle large frames, thereby maintaining compatibility while enabling efficient jumbo frame transmission on supported networks
Solution Approach 2:
The patent segments large jumbo frame payloads into multiple standard-sized Ethernet frames when transmission over networks with smaller MTU is required. The JFT device divides the large data payload into manageable chunks that conform to standard Ethernet frame limits, allowing efficient transmission across heterogeneous network infrastructures
2Productivity
If applications are modified to be aware of different Ethernet frame sizes, then transmission efficiency can be optimized, but application complexity and modification requirements increase
Solution Approach 1:
The JFT device acts as a transparent intermediary that handles frame size translation without requiring application modifications. Applications continue to send and receive data using standard Ethernet frames, while the JFT device automatically manages the conversion to and from jumbo frames, preserving application simplicity while achieving transmission efficiency
Solution Approach 2:
The system enables self-service by allowing the network infrastructure (JFT device) to automatically handle frame size adaptations without application involvement. The translation mechanism operates autonomously at the network layer, freeing applications from the burden of frame size management while maintaining optimal transmission performance
3Productivity
If jumbo frames are implemented directly over existing network stacks, then fewer packets are needed for large data transfer, but compatibility with networks having different MTU sizes is lost
Solution Approach 1:
The JFT device segments jumbo frames into standard-sized Ethernet frames for transmission over networks with smaller MTU. This segmentation allows the system to maintain high transmission efficiency on jumbo-capable networks while ensuring broad compatibility with standard Ethernet networks that cannot handle large frames
Solution Approach 2:
The system dynamically adapts frame sizes based on the capabilities of the receiving network. The JFT device determines the appropriate frame size to use based on the destination network's MTU, automatically switching between jumbo frames and standard frames to optimize both efficiency and compatibility across heterogeneous network environments
Data Source
AI summary
This disclosure is directed generally to systems and methods for implementation of Jumbo frames in an existing network stack. In some embodiments, a connection handler of a device receives data having a size greater than an Ethernet frame size. That data includes header data and payload data. The device partitions the data into segments including a first segment and a second segment. The first segment includes the header data and a first portion of the payload data, while the second segment includes a second portion of the payload data. The device stores the first and second segments in first and second network buffers, respectively, of a pool of network buffers. The device forms a packet chain of the first and second network buffers having a size greater than the Ethernet frame size. The device transmits the packet chain via a network connection.