Dual-Processor Architecture Offloads Network Traffic From Main CPUs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current communication processing platforms face challenges in efficiently handling simultaneous line rate bandwidth for secured data, voice, and video services without service degradation, leading to high CPU utilization and increased power consumption and heat generation.
Innovation Solution
A dual-processor architecture with specialized hardware engines for networking, security, and packet processing offloads intensive tasks from main CPUs, integrating these components into a single chip or package to free up resources for application and service tasks, utilizing flexible interconnects and hardware acceleration for efficient data processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If general-purpose main CPUs process all networking and data communication tasks, then networking performance is maintained, but application and service performance deteriorates due to CPU resource exhaustion
Solution Approach 1:
The patent segments processing tasks by dividing the system into general-purpose main CPUs for application/services and specialized data processing units for networking tasks. This segmentation allows each component to handle its designated workload independently, preventing resource exhaustion and maintaining both networking and application performance.
Solution Approach 2:
The patent introduces an intermediary data processing unit that acts as a mediator between network interfaces and main CPUs. This intermediary handles intensive networking tasks (packet routing, protocol processing, firewall operations) and communicates with main CPUs only when high-level application logic is required, reducing main CPU utilization while maintaining networking functionality.
2Productivity
If main CPUs handle intensive networking tasks, then data communication functions are performed, but power consumption and heat generation increase
Solution Approach 1:
The patent introduces specialized data processing units as intermediaries that handle intensive networking tasks. These units are optimized for specific data communication functions and consume less power than general-purpose main CPUs when performing repetitive networking operations, thereby reducing overall power consumption while maintaining data communication productivity.
Solution Approach 2:
The patent replaces the mechanical approach of using general-purpose CPUs for all tasks with specialized hardware accelerators or dedicated processing units for networking functions. This substitution allows efficient handling of data communication tasks through purpose-built circuitry that consumes less power than software-based processing on main CPUs.
3Speed
If more processing resources are allocated to networking tasks, then network bandwidth is maintained, but resources for application tasks are reduced
Solution Approach 1:
The patent segments processing resources into dedicated networking units and application processing units. This segmentation enables the system to maintain high network bandwidth through specialized data processing units while simultaneously preserving application processing capacity on main CPUs, as each segment operates independently with its own resource allocation.
Solution Approach 2:
The patent creates a multi-functional architecture where specialized data processing units handle networking tasks and main CPUs handle application services. This universal design allows the system to perform multiple functions simultaneously without resource contention, maintaining both network bandwidth and application processing capacity.
Data Source
AI summary
Communication traffic processing architectures and methods are disclosed. Processing load on main Central Processing Units (CPUs) can be alleviated by offloading data processing tasks to separate hardware. In one implementation, a processing architecture includes a main processor configured to execute a first portion of a driver software to perform protocol control and management task associated with control or management packets in a packet-based protocol according to which packets are received from a device, an offload processor configured to execute a second portion of the driver software to perform data processing task for data packets received according to the packet-based protocol, an interface to enable communication with the device, and an interconnect coupled to the main processor, to the offload subsystem, and to the interface.