L3 Switching Module Device Level Aggregation for FIB CAM Scaling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional data center network architectures face challenges in supporting a large number of hosts due to increased latency, complexity, and cost associated with growing Forwarding Information Base (FIB) and next-hop table sizes, especially when Virtual Local Area Networks (VLANs) span across multiple switching devices.
Innovation Solution
Implementing an IPv4 L3 switching module that uses subnet prefixes and device indices to aggregate forwarding information, reducing the size of the FIB CAM table by allocating bits for subnet prefixes, device indices, and host identifiers within IPv4 addresses, allowing for device-level aggregation and efficient packet forwarding across multiple switching devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If the number of hosts connected by switching systems increases to support more VMs, then the network capacity and host support capability improve, but the FIB CAM table size and next-hop table size increase dramatically, leading to increased cost and complexity
Solution Approach 1:
The patent segments the traditional monolithic FIB CAM table into multiple distributed FIB CAM tables across different switching devices. Each switching device maintains a portion of the forwarding information, eliminating the need for a single large table. This is achieved through distributed routing protocols that enable each device to independently manage its own FIB CAM entries based on local routing information.
Solution Approach 2:
The patent introduces a new dimension to forwarding by implementing two-stage forwarding: first-stage forwarding based on subnet prefix matching at the ingress switching device, and second-stage forwarding based on host-specific routing information. This dimensional approach allows the system to scale horizontally across multiple devices rather than vertically increasing a single table's capacity.
2Device complexity
If subnet prefixes are used as aggregation means, then FIB CAM table size is reduced, but forwarding cannot be performed when VLAN spans across different SDs because there is no conclusion on to which egress SD a packet should be forwarded
Solution Approach 1:
The patent introduces an intermediary mechanism in the form of a distributed routing protocol that operates between switching devices. This protocol exchanges routing information and enables forwarding decisions across device boundaries. The intermediary layer resolves the conflict between aggregation (smaller tables) and adaptability (VLAN spanning) by providing a coordination mechanism that allows subnet prefix aggregation while maintaining awareness of cross-device VLAN topology.
Solution Approach 2:
The patent implements dynamic forwarding capabilities where the system can adapt forwarding behavior based on the specific routing scenario. When VLANs span across devices, the distributed routing protocol dynamically provides the necessary forwarding information. This dynamic approach allows the system to maintain small static FIB CAM tables while achieving flexible adaptive forwarding for complex VLAN configurations.
3Ease of operation
If traditional three-tier network architecture is used, then network management is simplified, but as number of VMs grows, more switches and routers have to be added, causing dramatic increase on latency, complexity and cost
Solution Approach 1:
The patent extracts the routing functionality from the traditional three-tier architecture and distributes it across all switching devices in the network. By taking out the centralized routing control and distributing it to edge devices, the system eliminates the need for additional core switches and routers that would be required in a traditional scaling scenario. This extraction reduces both the number of devices and the latency associated with multiple forwarding hops.
Data Source
AI summary
Systems and methods are provided to support a large number of hosts while keeping the size of the FIB CAM low. Higher scalability may be achieved by consuming less FIB CAM space as hosts are added to a network. An L3 switching module may be provided capable of supporting a large number of virtual machines and/or hosts while maintaining a small FIB CAM size on a multiple device virtual switching system with device level aggregation.