GPU Packet Scheduling for CPU-Free 5G Transmission
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data transmission methods in 5G networks are bottlenecked by the limited processing capabilities of network interface cards, leading to computational intensity and degraded system performance due to CPU workload in preparing packet descriptors and scheduling transmission times.
Innovation Solution
Shift packet preparation, scheduling, and transmission responsibilities from the CPU to GPUs, utilizing their parallel processing capabilities to prepare and transmit data packets with precise timing instructions, thereby reducing CPU workload.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If CPU prepares packet descriptors and schedules transmission times, then packet transmission can be performed, but CPU workload increases and system performance degrades
Solution Approach 1:
The patent extracts the packet descriptor preparation and transmission scheduling functions from the CPU and relocates them to the network interface card. This is achieved by enabling the network adapter to directly access packet data in system memory and construct packet descriptors autonomously based on configuration parameters, thereby reducing CPU workload and allowing the CPU to focus on higher-level processing tasks while maintaining high packet transmission rates
Solution Approach 2:
The network interface card is empowered to perform self-service by autonomously preparing packet descriptors and scheduling transmissions without requiring continuous CPU intervention. The NIC maintains its own transmission queue and uses hardware-based scheduling mechanisms to manage packet transmissions, enabling the system to achieve high throughput while minimizing CPU involvement in routine packet handling operations
2Productivity
If network interface card processes packets, then transmission can occur, but processing capabilities are limited
Solution Approach 1:
The patent introduces a hybrid architecture where the network interface card acts as an intermediary between the CPU and the packet transmission process. The NIC includes dedicated hardware components for packet descriptor preparation and transmission scheduling, while also maintaining the ability to interact with the CPU for complex operations. This intermediary structure allows the system to leverage both the high-speed hardware processing of the NIC and the flexible programmable capabilities of the CPU
Data Source
AI summary
A computer based system and method for sending data packets over a data network may include: preparing data packets and packet descriptors on one or more graphical processing units (GPUs); associating packets with a packet descriptor, which may determine a desired transmission time of the packets associated with that descriptor; receiving an indication of a clock time; and physically transmitting packets via an output interface, at a clock time corresponding to the desired transmission time. A computer based system and method for GPU-initiated communication over a 5G data network may include allocating one or more memory buffers in GPU memory; performing at least one 5G signal processing procedure by a GPU; preparing descriptors for a plurality of packets, where each packet includes allocated memory buffers, and where the descriptors provide scheduling instructions for the packets; and triggering the sending of packets over the network based on prepared descriptors.