Distributed Semantic Network for Granular Resource Locking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In complex distributed systems, existing resource reservation techniques are not scalable and lead to inefficient handling of workflows due to the need to lock and update the state of many objects for a single operation, especially in dynamic environments with thousands of interconnected objects.
Innovation Solution
A distributed semantic network is used to manage resource reservations, where a set of facts and rules representing relationships between resources allows for concurrent access by obtaining and synchronizing knowledge across multiple instances, reducing the need to update the state of all linked and associated objects, and enabling granular locking and unlocking of resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the cloud management software reserves the entire system for each workflow using system-wide reservation, then resource conflicts are prevented, but the system becomes non-scalable and workflows are handled inefficiently
Solution Approach 1:
The patent segments the system from a monolithic system-wide reservation approach into granular object-level reservations. Each object (virtual machine, network device, storage device) can be independently reserved, allowing multiple workflows to access different objects concurrently. This segmentation enables parallel workflow execution while maintaining resource conflict prevention through fine-grained locking mechanisms.
2Reliability
If the cloud management software locks an object and all its child objects, then resource constraints are properly enforced, but the number of objects to update becomes very large
Solution Approach 1:
The patent applies local quality by implementing selective locking where only the specific object and its directly affected child objects are locked,而非 all objects in the system. The system determines the scope of locking based on the workflow's actual resource requirements, updating only the minimal necessary set of objects. This reduces the complexity of state management while maintaining proper constraint enforcement.
3Reliability
If the system updates the state of many objects for a single operation, then resource consistency is maintained, but the maximum scale of the system is limited
Solution Approach 1:
The patent implements dynamic object state management where the locking and updating scope adapts to each workflow's specific needs. The system dynamically determines which objects require locking based on the workflow definition and resource dependencies, rather than using a static all-or-nothing approach. This dynamic behavior maintains resource consistency for affected objects while enabling the system to scale to handle thousands of objects across multiple concurrent workflows.
Data Source
AI summary
An example method of reserving a resource of virtualized infrastructure in a data center on behalf of a client includes: obtaining, by a distributed semantic network, a set of facts corresponding to resources in the virtualized infrastructure and a set of rules corresponding to relationships between the resources; receiving, at the first semantic network instance, a first reservation request for a first resource of the virtualized infrastructure from a first client, wherein the first reservation request comprises a first rule specifying a requested exclusive lock on the first resource; passing the first rule from the first semantic network instance to the second semantic network instance; receiving an acknowledgement from the second semantic network instance in response to passing the first rule; and sending an acknowledgement to the first client that the first rule specifying the requested exclusive lock on the first resource has been created.