Replication Path Determination Across Failure Domains
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Computer networks and cloud computing environments face challenges in maintaining resiliency due to resources being associated with different failure domains, where a single point of failure can lead to the failure of all resources within that domain, disrupting service availability.
Innovation Solution
The solution involves determining and managing replication paths across multiple failure domains to ensure that resource consumers remain accessible even if multiple failure domains fail, by ordering resources into sequences and modifying replication paths to avoid loops that would reduce resiliency levels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If resources are replicated across multiple failure domains to improve resiliency, then service availability is improved, but system complexity increases
Solution Approach 1:
The system segments resources into distinct failure domains and creates ordered replication paths that traverse multiple domains. Each resource is assigned to a specific failure domain, and replication paths are constructed to sequentially pass through resources in different failure domains, ensuring that replicas are distributed across domain boundaries while maintaining manageable structural organization.
Solution Approach 2:
The patent introduces the dimension of ordered replication paths that traverse failure domains in sequence. Instead of simple one-to-one replication, the system creates multi-dimensional replication relationships where a single resource can replicate across multiple failure domains through an ordered sequence, adding structural dimensionality to the replication architecture.
2Reliability
If replication paths are extended to cover more failure domains to improve resiliency, then service availability is improved, but the risk of loops increases
Solution Approach 1:
The system incorporates feedback mechanisms that monitor the construction of replication paths and detect potential loops. When a replication path is being established, the system continuously checks whether adding the next resource would create a loop, and if so, it adjusts the path selection to avoid the loop while still maintaining resiliency objectives.
Solution Approach 2:
The patent performs preliminary loop detection and path validation before finalizing replication paths. By checking for potential loops in advance during the path construction phase, the system prevents loop formation before it occurs, rather than having to detect and resolve loops after they have been created.
Data Source
AI summary
Techniques for determining one or more replication paths for resources in different failure domains, while maintaining a target resiliency level, are disclosed. A replication path is a sequence of at least a subset of the set of resources. Based on the sequence, a resource is selected for replicating at least a portion of the resource consumers corresponding to each of the subset of resources. A replication path may be determined by adding a resource to the replication path and/or replacing one resource with another resource to modify the replication path. The modified replication path maintains the target resiliency level if the modified replication path does not include any loop of a length less than or equal to the target resiliency level.