Fabric Extender Multi-Destination Scale via Elected Forwarder
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data center network architectures face limitations in scaling multi-destination capabilities, particularly in supporting a high number of multicast groups and VLANs, due to the division of multi-destination indices between peer switches, which reduces the supported number of multicast groups and VLANs on fabric extenders (FEX) interfaces.
Innovation Solution
The proposed solution involves configuring a pair of switches with a virtual PortChannel (vPC) in an active/active topology, where one switch acts as the elected forwarder and the other as a standby, allowing the elected forwarder to program all multi-destination VIFs on the FEX, while the standby switch self-manages VIFs independently without dependency on the elected forwarder, enabling seamless failover and doubling the scale of supported multi-destination groups without increasing vPC peer-link usage or fabric fail-over time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multi-destination indices are divided between peer switches in a vPC configuration, then switch redundancy is maintained, but the scale of supported multi-destination groups is reduced
Solution Approach 1:
The patent segments the vPC peer switches into two distinct roles: elected forwarder and standby forwarder. This segmentation allows one switch (elected forwarder) to handle all multi-destination VIFs while the other (standby) provides redundancy, thereby resolving the contradiction between maintaining redundancy and supporting full multi-destination scale.
Solution Approach 2:
The elected forwarder switch performs multiple functions: it acts as a vPC peer switch and simultaneously programs all multi-destination VIFs on the FEX. This multi-functionality allows the system to maintain redundancy while concentrating multi-destination management capabilities in a single switch, doubling the supported scale compared to traditional divided approaches.
2Quantity of substance
If all multi-destination VIFs are programmed on a single switch, then the scale of supported groups increases, but the complexity of maintaining redundancy increases
Solution Approach 1:
The standby forwarder switch autonomously monitors the health of the elected forwarder and automatically transitions to the elected forwarder role upon detecting a failure. This self-service mechanism simplifies redundancy management by eliminating complex manual configuration and synchronization requirements while maintaining full multi-destination scale.
Solution Approach 2:
The system implements dynamic role assignment where switches can transition between standby and elected forwarder roles based on operational conditions. This dynamic behavior allows the system to maintain optimal performance and redundancy without requiring complex static configurations, thereby supporting increased multi-destination scale.
3Quantity of substance
If multi-destination scale is increased, then more multicast groups and VLANs are supported, but the vPC peer-link usage increases
Solution Approach 1:
The patent extracts the multi-destination VIF programming function from both peer switches and concentrates it solely on the elected forwarder. This extraction eliminates the need for synchronization traffic between peer switches for multi-destination VIFs, thereby increasing supported multicast groups without increasing vPC peer-link bandwidth consumption.
4Reliability
If multi-destination VIFs are synchronized between peer switches, then redundancy is improved, but the fabric fail-over time increases
Solution Approach 1:
The standby forwarder is pre-configured with the capability to assume the elected forwarder role, including having the necessary VIF information available through independent FEX programming. This preliminary preparation eliminates the need for time-consuming synchronization during failover, thereby achieving both redundancy and fast failover times.
Data Source
AI summary
An example method for increasing multi-destination scale in a network environment is provided and includes receiving a packet at a pair of switches comprising a first switch and a second switch in a network environment, where the first switch is configured as an elected forwarder, and the second switch is configured as a standby forwarder. A fabric extender (FEX) is connected to the first switch and the second switch with a virtual PortChannel. The method further includes forwarding the packet to a host connected to the FEX. The elected forwarder is configured with substantially all multi-destination virtual interfaces (VIFs) on the FEX in a forwarding table. The standby forwarder forwards the packet to the elected forwarder. In specific embodiments, the elected forwarder programs the FEX with the VIFs.