Causal Layer for AIOps Multi-Cloud Interoperability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current AIOps models are not interoperable across different cloud providers in multi-cloud computing systems due to inherent differences in cloud provider architectures and performance, leading to challenges in migrating applications and workloads while maintaining resilient, secure, and performant operations.
Innovation Solution
The introduction of a causal layer using chaos engineering to learn the inter-dependence between cloud providers and monitoring data, enabling AIOps model inter-operability without the need for fault injections or retraining on new datasets, by generating a causal model that maps configuration parameters to monitoring data and updating the AIOps model for seamless migration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If AIOps models are trained on cloud provider-specific data, then monitoring accuracy for that provider is improved, but interoperability across different cloud providers deteriorates
Solution Approach 1:
The patent introduces a causal layer as an intermediary component between the AIOps model and cloud provider-specific configuration parameters. This causal layer learns the inter-dependence relationships between different cloud providers' configurations and monitoring data, enabling the model to adapt to multiple cloud providers without retraining. The causal layer acts as a mediator that translates provider-specific parameters into a unified representation that the AIOps model can process, thereby maintaining monitoring accuracy across different providers while achieving interoperability.
2Adaptability or versatility
If AIOps models are retrained on new cloud provider datasets, then adaptation to the new provider is improved, but migration time and cost increase
Solution Approach 1:
The patent applies preliminary action by pre-training the causal layer on diverse cloud provider configuration data before deployment. This preliminary training enables the causal layer to learn generalizable inter-dependence relationships across different cloud providers. When migrating to a new cloud provider, the system does not need to retrain the entire AIOps model from scratch, as the causal layer has already learned to handle provider-specific variations. This significantly reduces migration time and computational costs while maintaining adaptation capability.
3Measurement precision
If extensive data collection is performed for each cloud provider, then model accuracy is improved, but data collection cost and time increase
Solution Approach 1:
The patent utilizes parameter changes by learning the inter-dependence relationships between cloud provider configuration parameters and monitoring data through the causal layer. Instead of collecting extensive data for each new cloud provider, the system leverages the learned parameter relationships to adapt to new providers with minimal data. The causal layer can infer how changes in provider-specific parameters affect monitoring data distributions, enabling accurate adaptation without extensive data collection for each provider.
Data Source
AI summary
Mechanisms are provided for migrating an application to a new cloud computing system. A causal model is generated based on configuration parameters for a first cloud computing system, monitoring data collected for an execution of the application in the first cloud computing system, and an inserted causal layer. Chaos engineering logic is executed on the causal model to perform a fault injection on the configuration parameters to emulate a second cloud computing system configuration. A mapping, by the causal layer, of the configuration parameters to the monitoring data is learned based on the fault injection. An artificial intelligence for information technology operations (AIOps) model is updated, based on the learned mapping of the causal layer, for monitoring the application execution in the new cloud computing system. The updated AIOps model is provided to an observability tool executing on the new cloud computing system.