Layer-2 TE Affinity Slicing in Layer-3 Bundled Links
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network slicing methods limit each layer-3 bundled link to a single network slice, restricting the flexibility of network slice division and preventing layer-2 member links from belonging to different network slices, which hinders traffic differentiation and isolation.
Innovation Solution
Configure TE affinity attributes for layer-2 member links within a layer-3 bundled link and flood these attributes using IGP, allowing each node to store layer-3 bundled link information, including specific TE affinity attributes of respective layer-2 member links, and create network slices using FA to enable multiple layer-2 member links to belong to different network slices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If each layer-3 bundled link is limited to a single network slice, then the structure is simple and easy to manage, but the flexibility of network slice division is restricted and layer-2 member links cannot belong to different network slices
Solution Approach 1:
The patent segments the association between layer-3 bundled links and network slices by introducing TE affinity attributes at the layer-2 member link level. Instead of binding a layer-3 bundled link to a single network slice, each layer-2 member link can be independently associated with different network slices through distinct TE affinity attributes, enabling fine-grained segmentation of network slice membership.
Solution Approach 2:
The patent applies local quality by allowing different layer-2 member links within the same layer-3 bundled link to have different TE affinity attributes. This enables each member link to be customized for specific network slice requirements, while the overall bundled link structure remains intact and manageable at the layer-3 level.
2Adaptability or versatility
If TE affinity attributes are configured for each layer-2 member link and flooded using IGP, then network slice division flexibility is enhanced, but the complexity of attribute management and flooding increases
Solution Approach 1:
The patent makes the TE affinity attribute mechanism universal by designing it to serve multiple purposes: it enables network slice identification, supports flexible association between layer-2 member links and network slices, and integrates with existing IGP flooding mechanisms. This multi-functionality reduces the need for separate management systems while enhancing flexibility.
Solution Approach 2:
The patent enables self-service by allowing network devices to automatically process and forward packets based on TE affinity attributes carried in packet headers. The IGP flooding mechanism automatically distributes TE affinity attribute information throughout the network, and devices autonomously make forwarding decisions without centralized control, reducing management overhead.
3Reliability
If layer-2 member links are allowed to belong to different network slices, then traffic differentiation and isolation are improved, but the difficulty of detecting and managing link attributes increases
Solution Approach 1:
The patent uses TE affinity attributes as symbolic 'colors' or identifiers that can be carried in packet headers and IGP messages. Each network slice can be associated with specific TE affinity attribute values, enabling easy detection and differentiation of traffic belonging to different network slices as packets traverse the network infrastructure.
Data Source
Figure 1~2B
Figure 2C~3
Figure 4~5
AI summary
A network slice creation method and apparatus, and a packet forwarding method and apparatus are provided. The network slice creation method includes: respectively configuring Traffic Engineering (TE) affinity attributes for layer-2 member links in a layer-3 bundled link, and flooding the TE affinity attributes of the layer-2 member links by using Interior Gateway Protocol (IGP); and then creating, according to the TE affinity attributes of the layer-2 member links and by using a Flex Algorithm (FA), network slices containing the layer-2 member links. In the embodiments of the present disclosure, the network slices containing the layer-2 member links are created according to the TE affinity attributes of the layer-2 member links and by using the FA, such that one or more layer-2 member links having the same TE affinity attribute can constitute a network slice. Therefore, it is possible to support a scenario in which various layer-2 member links under the layer-3 bundled link respectively belong to different network slices, and the division flexibility of the network slices can be improved.