Ethernet Bridge Path Selection Using Link Utilization Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current load distribution techniques in Ethernet networks with multiple equal cost paths lack a comprehensive view of traffic distribution, leading to inefficient bandwidth utilization and performance issues, especially in asymmetrical topologies, and require significant resource allocation and dilation factors to achieve reasonable network utilization.
Innovation Solution
The method involves an Ethernet Bridge with a topology database and a control processor that uses a shortest path search module to determine equal cost paths, a sorting module to rank paths based on link utilization, and a load distribution module to select subsets of paths that minimize standard deviation across the network, incorporating a tie-breaking process to resolve path selection and optimize link utilization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If algorithmic manipulation of node identifiers is used for path selection in 802.1aq, then pseudo-random path selection is achieved, but a significant dilation factor is required to even out link utilization
Solution Approach 1:
The patent implements feedback by collecting link utilization metrics from the network and using this information to dynamically adjust path selection. The system monitors actual link utilization and feeds this data back into the path computation algorithm, allowing it to adapt to changing network conditions and achieve better load distribution without requiring excessive dilation.
Solution Approach 2:
The patent transforms the static path selection based on algorithmic node identifier manipulation into a dynamic system that adapts to real-time network conditions. By continuously updating path selections based on current link utilization metrics, the system achieves more effective load distribution with a reduced dilation factor compared to static approaches.
2Productivity
If ECMP is used for load spreading, then load is divided pseudo-evenly across equal cost next hops, but the technique only considers local view and does not consider overall distribution of traffic in the network
Solution Approach 1:
The patent extends the path selection mechanism to serve multiple functions: it maintains the local ECMP functionality while simultaneously achieving global load balancing. The same path computation algorithm is used both for determining equal cost paths locally and for distributing traffic evenly across the entire network, eliminating the need for separate global optimization mechanisms.
Solution Approach 2:
The system collects global traffic distribution information through link utilization metrics and feeds this back into the path selection process. This feedback loop allows each router to make locally-based decisions that collectively achieve global load balancing, resolving the information loss problem inherent in purely local ECMP implementations.
3Productivity
If multiple VLANs are instantiated for each path permutation, then load distribution is achieved, but resource allocation increases significantly
Solution Approach 1:
The patent changes the parameters used for path selection from static node identifier-based algorithms to dynamic metrics-based selection. By using link utilization metrics as the primary selection criterion, the system achieves load distribution across fewer actual paths without requiring the creation of multiple VLANs for each path permutation, thereby reducing resource consumption.
Solution Approach 2:
The patent extracts the load distribution functionality from the VLAN instantiation mechanism and implements it through path selection algorithms. Instead of creating separate VLANs for load distribution, the system achieves the same effect by intelligently selecting which existing paths to use based on current network conditions, eliminating the need for excessive VLAN creation.
Data Source
AI summary
A method in an Ethernet Bridge for improved load distribution in an Ethernet network that includes the Ethernet Bridge including determining a first set of one or more shortest paths between each Ethernet Bridge pair in the Ethernet network, selecting at least a first shortest path, calculating a link utilization value for each link of the Ethernet network, determining a second set of one or more shortest paths between each Ethernet Bridge pair in the Ethernet network, generating a path utilization value for each shortest path, selecting a second shortest path on the basis of said path utilization value, whereby the selection of the second shortest in light of path utilization minimizes the standard deviation of load distribution across the entire Ethernet network.