Network Processor Single Point-of-Presence Multiprocessor Load Balancing

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current processor systems face bottlenecks in performance due to single network I/O interfaces and cache-contention, especially in advanced networking applications, and relying on Moore's law for improvements does not meet scalability needs, leading to increased complexity in managing multiple systems with separate network addresses.

Innovation Solution

A network element comprising a network processor that directs packets to specific traffic processors based on source and destination addresses associated with subscriber terminals, allowing for load balancing and maintaining a single IP/MAC address presence across multiple processors, thereby reducing network resource consumption and management burdens.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If multiple systems are introduced to achieve scalability, then processing power increases, but device complexity and management burden increase

Engineering Contradiction:
Improveprocessing powerVSAvoidmanagement complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The system segments processing functions across multiple independent processors while maintaining a unified network presence. Each processor can operate independently but collectively they present a single point of presence to the network, distributing workload while simplifying management through centralized addressing.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Multiple processors are merged into a single logical entity by assigning them a shared IP address and MAC address. This combining approach allows the system to leverage the processing power of multiple physical units while being managed and addressed as a single system, reducing management complexity.

Inventive Principle:
Principle #5Merging (Combining)

2Productivity

If multiple systems are introduced to achieve scalability, then processing power increases, but network resource consumption increases

Engineering Contradiction:
Improveprocessing powerVSAvoidnetwork addresses
Core Design Contradiction:
ProductivityVSQuantity of substance

Solution Approach 1:

Multiple processors share a single IP address and MAC address, merging their network identities. This approach allows the system to scale processing power across multiple physical units while consuming only one set of network addressing resources, significantly reducing network resource consumption compared to traditional multi-system deployments.

Inventive Principle:
Principle #5Merging (Combining)

3Ease of operation

If single processor systems are used, then management is simple, but processing power is limited

Engineering Contradiction:
Improvemanagement simplicityVSAvoidprocessing power
Core Design Contradiction:
Ease of operationVSProductivity

Solution Approach 1:

The system segments processing functions across multiple independent processors while maintaining unified management through shared network addressing. This segmentation enables parallel processing and increased throughput while preserving management simplicity through the single point of presence approach.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Multiple processors are configured to perform identical networking functions simultaneously, each capable of handling network packets independently. This multi-functionality allows the system to scale processing power while maintaining consistent management interfaces and single network identity.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentUS7826455B2Providing single point-of-presence across multiple processors
Publication Date: 2010.11.02 CISCO TECHNOLOGY INC
  • US7826455B2 patent drawing
  • US7826455B2 patent drawing
  • US7826455B2 patent drawing

AI summary

A method for providing single point-of-presence for a network element includes receiving a packet at a network processor, determining if the packet is to be directed to a particular one of a plurality of traffic processors if a source address of the packet is associated with a subscriber terminal, and determining if the packet is to be directed to the particular one of the plurality of traffic processors if a destination address of the packet is associated with the subscriber terminal. The method further includes distributing the packet to the particular one of the plurality of traffic processors.