Non-disruptive Network Policy Switching via Shadow Pipelines
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network configuration change methods often result in service disruptions, requiring prescheduled change windows and limiting dynamic or rapid implementation of policy changes.
Innovation Solution
Implementing a dual processing pipeline system where packets are processed according to an active policy while shadow policies are prepared in parallel, allowing for immediate policy switching without disrupting traffic by buffering and scheduling packets using shared resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If configuration changes are applied to network nodes to implement policy changes, then policy implementation capability is improved, but service disruption occurs and traffic processing is interrupted
Solution Approach 1:
The patent applies preliminary action by pre-configuring shadow processing pipelines with future policies before they become active. The shadow pipelines are prepared in advance with all necessary configuration changes, tested for validity, and staged for activation. This allows the system to have policy changes ready beforehand without applying them during traffic processing, thus avoiding service disruption while maintaining policy implementation capability.
Solution Approach 2:
The patent segments the processing pipeline into multiple independent components: active pipelines handling current traffic and shadow pipelines prepared for future policies. This segmentation allows configuration changes to be applied to shadow pipelines independently without affecting active pipelines. The system can switch between segments seamlessly, maintaining service continuity while enabling policy changes.
2Reliability
If configuration changes are applied during prescheduled change-windows, then service disruption is minimized, but dynamic or rapid policy implementation is limited
Solution Approach 1:
The system performs configuration changes in advance during off-peak periods or in preparation mode, staging multiple versions of processing pipelines with different policies. When a policy change is needed, the system can immediately activate a pre-prepared shadow pipeline without waiting for a prescheduled change-window, thus enabling rapid dynamic policy implementation while maintaining service continuity.
Solution Approach 2:
The patent introduces dynamic switching capability between active and shadow processing pipelines. The system can dynamically activate shadow pipelines with new policies at any time based on operational needs, rather than being constrained to prescheduled change-windows. This dynamic activation mechanism enables rapid policy implementation while maintaining service continuity through seamless switching.
3Adaptability or versatility
If multiple processing pipelines are maintained for different policies, then policy switching capability is improved, but device complexity increases
Solution Approach 1:
The patent uses shadow copies of processing pipelines that replicate the structure and functionality of active pipelines. These shadow pipelines are configured with future policies but remain inactive until needed. The copying approach allows the system to maintain multiple policy versions without requiring fundamentally different processing architectures, thus improving policy switching capability while managing device complexity through structured replication.
Solution Approach 2:
The patent designs processing pipelines with universal structures that can handle multiple policies through configuration rather than structural variation. The same pipeline framework can be configured with different policies by loading appropriate configuration sets into shadow pipelines. This multi-functionality approach reduces device complexity by using a unified pipeline architecture that serves multiple policy implementations.
Data Source
AI summary
Techniques for non-disruptive configuration changes are provided. A packet is received at a network device, and the packet is buffered in a common pool shared by a first processing pipeline and a second processing pipeline, where the first processing pipeline corresponds to a first policy and the second processing pipeline corresponds to a second policy. A first copy of a packet descriptor for the packet is queued in a first scheduler based on processing the first copy of the packet descriptor with the first processing pipeline. A second copy of the packet descriptor is queued in a second scheduler associated based on processing the second copy of the packet descriptor with the second processing pipeline. Upon determining that the first policy is currently active on the network device, the first copy of the packet descriptor is dequeued from the first scheduler.


