Service Model for Anomaly Detection in Shared Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Managing complex relationships between customers, services, and computing resources in a remote network management platform is challenging, leading to difficulties in tracking service dependencies and identifying impacts of resource failures, which can result in delayed recovery for customers.
Innovation Solution
A service model is implemented within the remote network management platform to capture and manage relationships between customers, services, and computing resources, using persistent storage and processors to determine affected customers and resources during anomalies or infrastructure updates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a managed network maintains hundreds or thousands of computing resources to support various services, then the service capacity and functionality are improved, but the complexity of tracking relationships between customers, services, and computing resources increases
Solution Approach 1:
The patent introduces a service model as an intermediary data structure that captures and manages the relationships between customers, services, and computing resources. This service model acts as a mediator layer that simplifies the complex many-to-many relationships by providing a structured representation that can be queried and analyzed, thereby reducing the complexity of tracking relationships while maintaining the ability to support numerous computing resources and services
2Adaptability or versatility
If the managed network expands the number of services and computing resources, then the functionality and service offerings are improved, but the difficulty of identifying affected customers during resource failures increases
Solution Approach 1:
The service model serves as an intermediary that explicitly captures the relationships between computing resources and affected customers. When a resource failure occurs, this structured model enables rapid querying to identify which customers are impacted, transforming a previously difficult detection problem into a straightforward lookup operation based on the pre-established relationships in the service model
3Ease of operation
If the managed network lacks real-time visibility into service dependencies, then the operational simplicity is maintained, but the recovery time for affected customers increases to days or weeks
Solution Approach 1:
The patent implements preliminary action by pre-establishing and maintaining the service model that captures all relationships between customers, services, and computing resources before failures occur. This advance preparation enables immediate identification of affected customers when failures happen, reducing recovery time from days or weeks to near-real-time without significantly complicating ongoing operations
Data Source
AI summary
A computational instance of a remote network management platform may be dedicated to a managed network. The computational instance may include persistent storage that contains: (i) mappings between end-user networks, services available to the end-user networks, and allocation identifiers, (ii) mappings between end-user networks, computing resources allocated to the end-user networks, and resource identifiers, and (iii) mappings between the respective allocation identifiers and the respective resource identifiers. Using the mappings, the computational instance may able to (i) determine that a particular computing resource is exhibiting an anomaly, (ii) determine a resource identifier associated with the particular computing resource, (iii) determine a allocation identifier based on the resource identifier, (iv) determine a particular end-user network associated with the allocation identifier; and (v) provide, to the managed network, an indication that the particular end-user network is potentially impacted by the anomaly.