Distributed Quota Management for Cloud Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Managing resource quotas in a cloud computing system is complex due to the need for frequent communication between a central quota manager and framework instances, which can lead to a single point of failure and increased communication overhead.
Innovation Solution
Implementing respective quota managers for framework instances that negotiate amongst themselves to manage resource usage, allowing each instance to reallocate sub-quotas when reaching a threshold, thereby minimizing communication and avoiding a single point of failure.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a central quota manager is used to manage resource quotas, then centralized control is achieved, but communication overhead increases and a single point of failure is created
Solution Approach 1:
The patent divides the centralized quota manager into multiple distributed quota manager instances, each managing a subset of framework instances. This segmentation eliminates the single point of failure by distributing the management function across multiple independent instances, while also reducing communication overhead by localizing quota management decisions.
Solution Approach 2:
The patent introduces a message broker as an intermediary component that enables asynchronous communication between quota manager instances and framework instances. This mediator decouples the communication channels, reducing direct communication overhead while maintaining system coordination through standardized message passing.
2Productivity
If frequent communication occurs between quota manager and framework instances, then real-time quota management is achieved, but network bandwidth consumption increases
Solution Approach 1:
The patent implements event-driven periodic communication where quota manager instances and framework instances communicate only when specific events occur (e.g., quota threshold violations, reallocation requests). This reduces network bandwidth consumption by eliminating continuous polling while maintaining real-time responsiveness to actual quota conditions.
Solution Approach 2:
The patent enables framework instances to autonomously monitor their own resource usage and self-report when they need quota reallocation, rather than requiring continuous external monitoring. This self-service approach reduces communication overhead by triggering messages only when necessary, improving efficiency while minimizing network bandwidth consumption.
3Measurement precision
If detailed quota tracking is implemented, then accurate resource management is achieved, but processing requirements increase
Solution Approach 1:
The patent implements threshold-based quota tracking where detailed monitoring is applied only when resource usage approaches predefined thresholds. Below these thresholds, simplified tracking is sufficient, reducing processing requirements while maintaining accurate measurement precision when it matters most for quota management decisions.
Data Source
AI summary
Techniques are provided for automated employment of respective quota managers for framework instances, where the respective quota managers can negotiate amongst themselves to manage usage of a resource of a shared computing system in relation to a quota for the resource for a tenant of the shared computing system. This can allow tenants to share their quota among multiple frameworks, enable quota exchange between multiple frameworks, and choose a quota with a minimum costs, and thus maximize savings.