Forwarding Table Atomic Updates During Live Packet Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Networking devices face challenges in efficiently updating forwarding tables during live packet stream processing without disrupting network packet processing, due to limitations such as heat generation and resource constraints, which can lead to fragmentation and inefficiencies in handling high workloads.
Innovation Solution
Implementing atomic updates to forwarding tables by allocating reserved entries for updating operations, copying entries to these reserved areas, and modifying access pointers to ensure continuous packet processing, thereby defragmenting and reorganizing entries without stalling the network flow.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If forwarding table updates are performed during live packet processing, then network adaptability and routing flexibility are improved, but processing stability and consistency may deteriorate due to potential fragmentation and interruptions
Solution Approach 1:
The forwarding table is segmented into multiple regions or structures that can be updated independently. The patent divides the forwarding table into different segments that can be modified separately during live operation, allowing routing flexibility to improve while maintaining processing stability through independent segment management.
Solution Approach 2:
The patent prepares forwarding table updates in advance by allocating reserved entries and pre-configuring update structures before they are needed. This preliminary action ensures that updates can be applied atomically without disrupting live packet processing, maintaining both adaptability and stability.
2Productivity
If networking device capacity is increased to handle high workloads, then processing speed and throughput are improved, but heat generation and power consumption worsen
Solution Approach 1:
The patent implements dynamic forwarding table updates that adapt to current workload conditions. By allowing incremental updates during live operation rather than requiring full table rebuilds, the system maintains high processing speeds while reducing the power consumption and heat generation associated with frequent full-table updates.
Solution Approach 2:
The patent ensures continuous packet processing by implementing update mechanisms that do not interrupt the data plane. Forwarding table updates occur in the control plane while packet forwarding continues uninterrupted, maintaining productivity without the energy costs of stopping and restarting processing.
3Productivity
If forwarding table updates are performed during live processing, then resource utilization is improved, but update complexity and potential for errors increase
Solution Approach 1:
The patent introduces intermediary structures such as reserved entries and update buffers that mediate between the control plane update operations and the data plane forwarding operations. These intermediaries simplify the update process by providing standardized interfaces and reducing the complexity of coordinating updates with live processing.
Solution Approach 2:
The patent uses copying mechanisms where updated forwarding entries are prepared in reserved regions and then copied or swapped into the active forwarding table. This copying approach reduces update complexity by separating the update preparation from the actual table modification, allowing atomic updates without complex coordination.
4Reliability
If atomic updates are implemented to prevent fragmentation, then processing reliability is improved, but update time and latency may worsen
Solution Approach 1:
The patent segments the forwarding table into multiple independently updatable regions. By dividing the table into segments that can be updated separately and atomically, the system maintains high reliability through consistent updates while reducing overall update time by parallelizing updates across multiple segments rather than requiring a single long atomic operation.
Data Source
AI summary
A packet processor or packet processing pipeline may implement forwarding table updates during live packet stream processing. Updates may be updates to add or remove entries from group of entries in a forwarding table or may be updates to defragment the available entries in a chain of entries maintained in a forwarding table. Reserved entries may be allocated for moving entries as part of updates so that migrations of the entries from one location to another may be performed atomically.


