IP Fabric Bandwidth Distribution via BGP Path Attributes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data center networks lack an efficient mechanism to distribute network traffic based on varying link bandwidths, leading to suboptimal bandwidth utilization and potential congestion, as conventional BGP routing protocols do not effectively propagate link bandwidth information across multi-hop paths.
Innovation Solution
The implementation of a mechanism that aggregates and propagates link bandwidth information using a Path Bandwidth extended community attribute within BGP, allowing network devices to calculate and advertise end-to-end path bandwidth ratios, enabling intelligent data forwarding decisions based on available upstream bandwidths.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional BGP routing protocols are used for traffic distribution, then routing functionality is provided, but link bandwidth information is not effectively propagated across multi-hop paths leading to suboptimal bandwidth utilization
Solution Approach 1:
The patent introduces a Path Bandwidth extended community attribute as an intermediary mechanism that carries link bandwidth information through BGP routing updates. This attribute acts as a mediator between the routing protocol and bandwidth management, enabling multi-hop propagation of bandwidth data without modifying the core BGP protocol functionality.
Solution Approach 2:
The patent segments the bandwidth information into discrete Path Bandwidth attributes that can be independently attached to BGP routing updates. This allows bandwidth information to be carried separately from routing path information, enabling selective propagation and utilization of bandwidth data across the network.
2Productivity
If traffic is distributed without considering upstream bandwidth availability, then simple forwarding decisions are made, but congestion occurs and network performance deteriorates
Solution Approach 1:
The patent implements preliminary action by having network devices calculate and advertise available upstream bandwidth to downstream devices before traffic forwarding decisions are made. This advance preparation of bandwidth information enables informed routing decisions without adding complexity to the actual forwarding operation.
Solution Approach 2:
The patent establishes a feedback mechanism where network devices advertise their available upstream bandwidth back to downstream devices through BGP updates. This feedback loop allows the network to dynamically adjust traffic distribution based on actual bandwidth availability, preventing congestion while maintaining relatively simple forwarding logic.
3Adaptability or versatility
If devices are upgraded incrementally in the data center, then cost-effective deployment is achieved, but various devices have different capabilities creating bandwidth distribution challenges
Solution Approach 1:
The patent applies local quality by allowing each network device to advertise its specific available upstream bandwidth capacity through the Path Bandwidth attribute. This enables the network to recognize and utilize the different capabilities of incrementally upgraded devices, with each device contributing its local bandwidth characteristics to the overall network optimization.
Data Source
AI summary
In one example, a plurality of network devices forming an Internet protocol (IP) fabric includes first, second, third, and fourth network devices. The first network device includes a plurality of network interfaces communicatively coupled to at least the third and fourth network devices of the plurality of network devices, which are between the first network device and the second network device. The first network device also includes one or more hardware-based processors configured to determine available bandwidths for the third network device and the fourth network device toward the second network device, determine a ratio between the available bandwidths for the third and fourth network devices, and forward data (e.g., packets or bytes) toward the second network device such that a ratio between amounts of the data forwarded to the third and fourth network devices corresponds to the ratio between the available bandwidths.