Scaling Framework for Cloud Query-Based Resource Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current cloud management software lacks user control over scaling and alerting decision logic, fails to support custom metric sources, and does not allow for customizable eviction policies, leading to inefficient resource management and suboptimal scaling decisions.
Innovation Solution
A scalable system with a scaling engine that includes a scaling master and agents, connected through a structured query language interface, allowing users to specify and execute scaling and alerting queries, form decisions based on aggregated results, and manage processing elements effectively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If scaling and alerting decision logic is shipped as a built-in feature, then the system provides basic scalability, but users have limited control over the logic and cannot customize complex scaling conditions
Solution Approach 1:
The patent extracts the scaling and alerting decision logic from the built-in feature and places it in an external, user-configurable location. Users can now specify custom scaling conditions, metrics, and policies outside the core system, allowing full control over scaling behavior without modifying the core product architecture.
Solution Approach 2:
The system transitions from static, pre-defined scaling logic to dynamic, user-configurable scaling policies. Users can dynamically adjust scaling conditions, thresholds, and behaviors based on their specific needs, making the system adaptable to various workloads and requirements.
2Adaptability or versatility
If simple UI-based scaling solutions are provided, then the system is easy to operate, but it cannot evaluate complex scaling conditions
Solution Approach 1:
The patent segments the scaling configuration into distinct, manageable components: metric sources, scaling policies, alerting rules, and decision logic. Each component can be configured and managed independently, allowing users to build complex scaling conditions through modular configuration rather than monolithic settings.
Solution Approach 2:
The system introduces an intermediary configuration layer that sits between the simple UI and the complex scaling logic. This intermediary allows users to define custom metrics, policies, and conditions using a structured approach, bridging the gap between simplicity and complexity.
3Adaptability or versatility
If fixed eviction policies are used, then the system is simple to implement, but processing elements that are not the best suited for eviction may be removed
Solution Approach 1:
The eviction policy transitions from a fixed, static rule to a dynamic, customizable policy. Users can define eviction criteria based on multiple factors such as resource usage patterns, workload characteristics, and performance metrics, allowing the system to intelligently select which processing elements to evict based on current conditions.
Solution Approach 2:
The system allows users to change multiple parameters of the eviction policy, including thresholds, time windows, metric weights, and priority levels. By adjusting these parameters, users can fine-tune the eviction behavior to match their specific workload requirements and performance goals.
4Adaptability or versatility
If cloud management software provides little support for custom metric sources, then the system is simple to maintain, but it cannot support customized scalability and alerting requirements
Solution Approach 1:
The patent implements a universal metric source interface that can work with multiple different data sources and formats. The system provides a standardized way to define and query custom metrics while supporting various underlying sources, making the system versatile without requiring separate integration logic for each metric type.
Solution Approach 2:
The metric source functionality is segmented into independent, configurable components. Users can define custom metrics as separate entities with their own data sources, calculation methods, and aggregation rules, allowing flexible metric definition without complicating the core system architecture.
Data Source
AI summary
Certain example embodiments described herein relate to techniques for scaling processing systems. For instance, in certain example embodiments, a scaling master receives a user-specified scaling and/or alerting query. The scaling and/or alerting query is transmitted to scaling agents at respective ones of the processing elements, the transmission using the connection layer. Each of the scaling agents executes the at least one scaling and/or alerting query. Each of the scaling agents reports a result of the executing to the scaling master using the connection layer. The scaling master forms a scaling decision based on the result reported by each of the scaling agents.