Switch Configuration Viability Prediction via SSM
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current switch operating systems in massively scalable data centers lack the ability to predict whether a proposed configuration or interface event can be successfully accepted by the hardware, leading to potential system failures and inconsistencies, especially in virtual Port-channel scenarios where configurations must be synchronized across multiple switches.
Innovation Solution
A distributed system scale management (SSM) architecture that tracks resource usage and capacity across multiple compute nodes, allowing for dynamic load shaping and prediction of configuration viability by using a Verify/Reserve/Commit Protocol to ensure successful configuration application across switches.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If the controller dumps the entire configuration onto the switch, then the configuration process is simplified, but the switch cannot predict whether the configuration will be successfully accepted by the hardware
Solution Approach 1:
The patent applies preliminary action by performing resource availability checking and configuration validation before the actual configuration is applied to the switch. The controller queries the switch about resource availability, and the switch provides predictions about whether the configuration will be successfully accepted by the hardware, allowing the controller to prepare and validate configurations in advance.
Solution Approach 2:
The patent implements feedback by having the switch provide information back to the controller about resource availability and configuration viability predictions. This feedback mechanism allows the controller to make informed decisions about configuration application, adjusting its approach based on the switch's predictions about hardware acceptance.
2Adaptability or versatility
If multiple interfaces generate concurrent resource-consuming events like LINK_UP, then the system responds dynamically to interface events, but unusual system load occurs due to bringing up a number of interfaces in a short span of time
Solution Approach 1:
The patent applies preliminary action by having the controller receive predictions from the switch about resource availability and configuration acceptance before actually applying configurations that would trigger interface events. This allows the controller to plan and stagger interface bring-up events to avoid concurrent resource-consuming operations.
Solution Approach 2:
The patent implements dynamics by enabling the system to adaptively manage interface event handling based on predicted resource availability. The controller can dynamically adjust the timing and sequencing of interface bring-up events based on feedback from the switch about current and future resource states, preventing system overload while maintaining responsiveness.
3Reliability
If the switch tracks resource usage and capacity across multiple compute nodes, then configuration viability can be predicted, but the system complexity increases
Solution Approach 1:
The patent applies the intermediary principle by introducing a prediction mechanism within the switch that acts as a mediator between the controller and the hardware resources. The switch's resource tracking and prediction capabilities serve as an intermediary layer that provides information about configuration viability without requiring the controller to directly manage complex resource tracking across multiple compute nodes.
Data Source
AI summary
Methods and systems for configuring a local switch and a remote switch configured as peers. The method includes: receiving a configuration for configuring the local switch and the remote switch; checking if resources are available for the configuration on the local switch; when resources are available on the local switch, reserving those resources; checking if resources are available for the configuration of the remote switch; when resources are available on the remote switch, reserving those resources; and applying the configuration to the local switch and the remote switch when resources are available on the local switch and the remote switch.