Distributed OS Coherency in Multi-Chassis Routers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
As networks become increasingly complex, managing state information across multiple components in multi-chassis routers becomes challenging, leading to inefficiencies and interruptions in packet forwarding due to redundant execution of operating systems and the need for redundant state information propagation.
Innovation Solution
A distributed operating system is implemented across computing nodes, allowing for dynamic resource expansion or contraction, robust state information convergence, and efficient synchronization using an object flooding protocol, which eliminates redundant execution and simplifies state information propagation through multicast delivery.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple routing engines each execute separate instances of the operating system to manage state information, then reliability is improved through failover capability, but device complexity increases and bandwidth requirements increase due to redundant state information propagation
Solution Approach 1:
Multiple routing engines share a single instance of the distributed operating system instead of each executing separate instances. The routing engines collaborate through the shared OS instance, which maintains unified state information that is propagated to all consumers, eliminating redundancy while maintaining failover capability through the shared architecture.
2Reliability
If standby routing engines relearn lost state information from each resource by power cycling, then reliability is maintained through state recovery, but loss of time occurs due to interruption in packet forwarding
Solution Approach 1:
State information is pre-synchronized and maintained across routing engines through the distributed operating system before failures occur. When a failure happens, the standby engine already possesses the necessary state information through prior synchronization, eliminating the need for time-consuming relearning and power cycling operations.
Solution Approach 2:
The distributed operating system acts as an intermediary that maintains and propagates state information to all routing engines. This intermediary mechanism ensures that state information is readily available to standby engines without requiring direct resource-by-resource relearning, thus reducing recovery time and avoiding packet forwarding interruptions.
3Adaptability or versatility
If each component in multi-chassis routers executes its own instance of the operating system, then adaptability is improved for managing diverse resources, but loss of information increases due to difficulties in maintaining coherent state information across nodes
Solution Approach 1:
A single instance of the distributed operating system serves multiple routing engines and resources simultaneously, providing universal state management capability. This unified instance maintains coherent state information that is accessible and applicable across all diverse resources, eliminating the information loss that would occur with multiple separate instances.
Solution Approach 2:
The distributed operating system implements feedback mechanisms through state update notification messages that propagate state information changes to all consumers. This feedback loop ensures that all routing engines receive and utilize the same state information, maintaining coherence across the multi-chassis router system while preserving adaptability to diverse resources.
4Productivity
If state information is cached within the operating system instance and update notifications are issued to consumers, then productivity is improved through efficient state access, but device complexity increases due to the propagation mechanism
Solution Approach 1:
The distributed operating system instance automatically manages state information caching and propagation without requiring complex external coordination mechanisms. The system self-services by maintaining the state cache and issuing update notifications to consumers as needed, improving productivity while keeping the propagation mechanism relatively simple through its self-managing nature.
Data Source
AI summary
In general, techniques are described for maintaining coherency in distributed operating systems for network devices. A network device comprising hardware computing nodes may be configured to perform the techniques. The hardware computing nodes may execute a distributed operating system. At least one the hardware computing nodes may determine whether one or more of the plurality of hardware computing nodes has failed and is no longer supporting execution of the distributed operating system, and determine whether remaining ones of the plurality of hardware computing nodes exceeds a quorum threshold. The at least one of the hardware computing nodes may further restart, when the remaining ones of the plurality of hardware computing nodes is less than the quorum threshold, the distributed operating system.