Hybrid Multicast Switching for MLID-Limited Network Routing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing multicast configurations in networks face challenges with slow and resource-intensive replication of data to multiple destination endpoints, particularly in large and unpredictable communication patterns, due to limited multicast local identifiers (MLIDs) and unpredictable communication patterns.
Innovation Solution
Implementing a hybrid-multicast operation using a bitmask representation in addition to MLIDs, allowing switches to dynamically select additional egress ports for data transmission, enhancing static multicast routing with dynamic routing capabilities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional multicast routing configuration is used in network switches, then data can be replicated to multiple destination endpoints, but the process is slow and cannot accommodate data-dependent communication patterns efficiently
Solution Approach 1:
The patent introduces dynamic multicast routing capabilities that allow network switches to adapt routing decisions based on data-dependent communication patterns. The system uses runtime information about destination endpoints and communication patterns to dynamically adjust multicast group configurations and routing paths, transforming the static multicast routing into a dynamic system that can respond to changing network conditions and workload requirements.
Solution Approach 2:
The patent changes key parameters of multicast routing by introducing new configuration options that allow switches to modify routing behavior based on workload characteristics. This includes changing how multicast groups are formed, how destination endpoints are selected, and how routing decisions are made, enabling the system to optimize for different communication patterns without requiring fundamental architectural changes.
2Quantity of substance
If more multicast local identifiers (MLIDs) are used to support more destination endpoints, then more endpoints can be addressed, but MLIDs are a limited resource
Solution Approach 1:
The patent makes MLIDs more universal by enabling a single MLID to serve multiple destination endpoints through dynamic group formation. Instead of requiring dedicated MLIDs for each endpoint or small groups, the system allows one MLID to dynamically represent larger multicast groups composed of multiple endpoints, thereby reducing the total number of MLIDs needed in the network while increasing the number of addressable endpoints.
Solution Approach 2:
The patent segments the multicast routing function into multiple components: MLID assignment, dynamic group formation, and runtime routing decisions. This segmentation allows the system to manage limited MLID resources more efficiently by separating the identification function from the routing function, enabling more endpoints to be reached through smarter grouping rather than through sheer numbers of identifiers.
3Device complexity
If static multicast routing configuration is used, then routing is simplified, but it cannot accommodate unpredictable communication patterns and large network topologies efficiently
Solution Approach 1:
The patent applies preliminary action by pre-configuring basic multicast routing structures and MLID assignments, but leaves the actual routing decisions to be made dynamically at runtime based on observed communication patterns. This allows the system to have simple initial configurations while still achieving adaptive behavior through runtime adjustments, balancing complexity and adaptability.
Data Source
Figure 1
Figure 2
Figure 3A
AI summary
Systems and methods herein are for multicast configurations in a network using a switch that can receive communication for a multicast group and that can transmit at least data from the communication to different destination nodes. The communication may include, in addition to the data, a multicast identifier associated with first egress ports of the switch and a bitmask representation associated with second egress ports of the switch. The transmission of the data can occur through the first egress ports and the second egress ports to reach the different destination nodes.