Composed Service Fault Isolation Using Resource Overlap Matrices
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In virtual clustered infrastructure, fault domains are not clearly defined, leading to potential failure of multiple virtual network servers due to the shared use of dynamically provisioned resources, which is not addressed by existing technologies.
Innovation Solution
The implementation of fault matrices for composed services and clusters to identify and minimize overlapping physical resources, ensuring fault isolation by minimizing shared resources among composed services.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If dynamically provisioned resources are shared among composed services to improve resource utilization, then resource efficiency is improved, but fault isolation deteriorates causing potential simultaneous failure of multiple services
Solution Approach 1:
The patent segments the fault domain by introducing a fault domain identifier (FDID) that groups composable resources into distinct fault domains. Services are assigned to specific fault domains, creating logical segmentation that prevents cascading failures across the entire system while maintaining resource sharing within domains. This resolves the contradiction by organizing shared resources into isolated failure zones.
Solution Approach 2:
The patent introduces an intermediary layer (the fault domain assignment mechanism and FDID metadata) between the composable resources and the composed services. This intermediary manages the mapping relationship, allowing resources to be dynamically shared while maintaining fault isolation through the intermediary's control over which services can access which resources within their assigned fault domain.
2Reliability
If fault matrices are built and searched to identify overlapping resources, then fault isolation is improved, but system complexity increases
Solution Approach 1:
The patent performs preliminary action by pre-building fault matrices that capture the relationships between composable resources and composed services before actual service deployment or failure scenarios occur. The fault matrices are constructed in advance, storing the mapping relationships that enable rapid fault isolation analysis without requiring complex real-time computations when failures happen.
Solution Approach 2:
The patent creates a simplified copy or representation of the complex resource-service relationships in the form of fault matrices. Instead of directly analyzing the complex dynamic resource provisioning system during failure events, the system uses these matrix representations to efficiently identify overlapping resources and potential failure paths, reducing the computational complexity of fault analysis.
Data Source
AI summary
A method for processing service fault assessment requests. The method includes: receiving a service fault assessment request concerning a new composed service; making a determination that the new composed service is already configured; based on the determination: building a service fault matrix for the new composed service; obtaining a cluster fault matrix for a composed service cluster to which the new composed service belongs; making a second determination, based on a search performed using the service fault matrix and the cluster fault matrix, that an overlapping composable resource has been identified; based on the second determination: generating a service fault assessment report including identification metadata associated with the overlapping composable resource; and providing the service fault assessment report in response to the service fault assessment request.