Distributed Bridging Fabric Synchronization via Control Plane
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In distributed packet-switched communication networks, achieving synchronized forwarding databases across multiple units to maintain consistent data traffic forwarding and minimize flooding, especially when units or links fail, is challenging due to independent learning and aging mechanisms.
Innovation Solution
The implementation of explicit synchronization through broadcasting maintenance packets for learning events and aging management across fabric ports, along with database resynchronization when units join or leave the fabric, ensures consistent forwarding database entries and minimizes traffic flooding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If independent learning and aging mechanisms are used in each unit, then each unit can operate autonomously, but forwarding databases become desynchronized causing inconsistent data forwarding and increased traffic flooding
Solution Approach 1:
The patent implements a feedback mechanism where the control plane monitors the state of forwarding databases across fabric units and initiates resynchronization when desynchronization is detected. This feedback loop ensures that autonomous units maintain synchronized forwarding databases, resolving the contradiction between autonomous operation and database synchronization.
Solution Approach 2:
The control plane acts as an intermediary between autonomous fabric units, coordinating their forwarding database operations. Through control plane messaging and resynchronization protocols, the intermediary ensures that independent units maintain consistent forwarding information without sacrificing their autonomous operational capability.
2Productivity
If distributed bridging is implemented across multiple units, then workload is shared and data forwarding performance is maximized, but maintaining a single forwarding database identity becomes complex
Solution Approach 1:
The patent segments the forwarding database functionality by separating it into data plane (forwarding information base) and control plane (resynchronization management). This segmentation allows multiple fabric units to share forwarding workload while the control plane handles the complexity of maintaining unified database identity, thus improving performance without overwhelming individual units.
Solution Approach 2:
The patent merges the forwarding database operations of multiple fabric units into a unified logical database through control plane coordination. While physically distributed across units for performance, the databases are logically combined through resynchronization mechanisms, presenting a single forwarding database identity to the network while maintaining distributed performance benefits.
3Loss of information
If explicit synchronization through maintenance packets is implemented, then forwarding databases remain synchronized, but network overhead increases
Solution Approach 1:
The patent implements periodic resynchronization actions triggered by control plane events rather than continuous maintenance packet exchange. This periodic approach synchronizes forwarding databases at appropriate intervals and events, maintaining consistency while minimizing the energy overhead associated with constant communication between fabric units.
Solution Approach 2:
The control plane performs self-service by autonomously detecting forwarding database desynchronization and initiating resynchronization without requiring continuous external maintenance packets. This self-service mechanism reduces network overhead by only communicating when necessary, while still ensuring forwarding database consistency across fabric units.
Data Source
AI summary
A network unit for use in a distributed bridging fabric has a multiplicity of user ports for the transmission of data frames to and from an external network and at least one fabric port for the transmission of frames between the network unit and another unit in the fabric. The network unit has a forwarding database for containing entries each including a media access control address, and a lookup engine organized for the insertion of an entry into the forwarding database when the network unit receives a data packet at a user port. The unit broadcasts via each fabric port of an “address added” message identifying a respective media access control address. The lookup engine also responds to such an “address added” message received from another unit to make a corresponding entry in the database and to annotate an entry to indicate activity of an address in response to an address which is already in the database. The network unit has an aging engine organized for the polling of entries in the database. The aging engine refreshes an entry for which the respective media access control address is annotated as active and is organized for the selective removal of inactive entries from the database.