Network Element Allocation for Computational Node Slices

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing data center networks face challenges in providing customized network resources to accommodate diverse application requirements, such as bandwidth-intensive and latency-sensitive applications, within a shared infrastructure.

Innovation Solution

A computer program product, system, and method for allocating dedicated network elements to slices of nodes in a network, allowing for customized network topologies and resource allocation to meet specific workload requirements, including switch routing rules to route packets through dedicated network elements.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If dedicated network elements are allocated to slices of nodes, then network performance and traffic isolation are improved, but network complexity and resource allocation difficulty increase

Engineering Contradiction:
Improvenetwork performanceVSAvoidnetwork complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The network is segmented into multiple slices, where each slice represents an isolated virtual network environment. Network elements are allocated to specific slices to provide dedicated paths for computational nodes within each slice, ensuring traffic isolation and improved reliability while maintaining manageable complexity through structured segmentation

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces a virtualization dimension by creating logical slices over the physical network infrastructure. This allows multiple isolated network environments to coexist on the same physical hardware, improving network performance and isolation without proportionally increasing physical device complexity

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Object-affected harmful factors

If dedicated network elements are allocated to slices of nodes, then traffic isolation between different applications is improved, but device complexity and routing configuration increase

Engineering Contradiction:
Improvetraffic interferenceVSAvoidrouting configuration
Core Design Contradiction:
Object-affected harmful factorsVSDevice complexity

Solution Approach 1:

Traffic isolation is achieved by segmenting the network into distinct slices, where each slice contains dedicated network elements and routing rules. This prevents traffic from different applications (bandwidth-intensive and latency-sensitive) from interfering with each other, while the segmented structure keeps routing configuration manageable through modular organization

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Each slice receives customized network resources and routing rules tailored to its specific requirements. Latency-sensitive applications get priority routing within their slice, while bandwidth-intensive applications receive adequate capacity allocation, providing local quality optimization without requiring global reconfiguration of the entire network

Inventive Principle:
Principle #3Local quality

Data Source

PatentUS12301465B2Allocating network elements to slices of nodes in a network
Publication Date: 2025.05.13 INTERNATIONAL BUSINESS MACHINE CORPORATION
  • US12301465B2 patent drawing
  • US12301465B2 patent drawing
  • US12301465B2 patent drawing

AI summary

Provided are a computer program product, system, and method for allocating network elements to slices of nodes in a network. Dedicated network elements are allocated to slices, defining sets of computational nodes, to provide dedicated paths in the network for the computational nodes in the slices to interconnect with other computational nodes in a same slice. The computational nodes connect to leaf switches and the leaf switches connect to spine switches. The spine switches interconnect the leaf switches. The dedicated network elements comprise at least one of a switch, an edge of at least one link, a link, and at least one queue of a port for a link. Switch routing rules are provided to at least one switch including the dedicated network elements to route packets from the computational nodes through the dedicated network elements allocated to the slices of the computational nodes.