Storage Server Resource Control via Selective Microservice Activation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing storage solutions waste resources and fail to meet performance requirements due to unnecessary execution of grouped tasks, especially under high-bandwidth workloads.
Innovation Solution
A resource manager logically decouples grouped services, suspending or activating individual services based on resource mappings and consumption models, using machine learning to optimize resource allocation and service initiation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If grouped services are executed together as a unit, then service reliability is improved, but resource consumption increases and latency worsens
Solution Approach 1:
The patent segments grouped services into individual microservices that can be independently activated. Instead of executing all services in a group together, the system identifies and activates only the specific microservices needed to handle each request, reducing unnecessary resource consumption while maintaining reliability through selective service execution.
Solution Approach 2:
The system dynamically determines which microservices to activate based on request characteristics and current system state. The service activation is not static but adapts to different workload conditions, allowing the system to optimize resource usage while ensuring appropriate services are available when needed.
2Reliability
If all services in a grouped service are activated by default, then service completeness is improved, but resource consumption increases
Solution Approach 1:
Instead of activating all services in a grouped service by default (excessive action), the system activates only the partial set of services that are actually needed to process each specific request. This avoids the waste of resources on unnecessary service executions while ensuring that sufficient services are activated to complete the required tasks.
3Ease of operation
If grouped tasks are executed together, then task coordination is simplified, but latency increases under high-bandwidth workloads
Solution Approach 1:
The patent segments grouped tasks into individual microservices that can be independently executed and managed. This segmentation allows the system to activate only the necessary microservices for each request, reducing the overall execution time and latency compared to waiting for all grouped tasks to complete, while still maintaining proper coordination through the microservice architecture.
4Productivity
If individual microservices are activated selectively, then resource consumption is reduced, but service coordination complexity increases
Solution Approach 1:
The system introduces an intermediary component (the service activation mechanism) that manages the coordination of selectively activated microservices. This intermediary handles the complexity of determining which services to activate and coordinating their execution, thereby reducing the apparent complexity for the rest of the system while enabling efficient resource usage through selective activation.
Data Source
AI summary
Aspects of the present disclosure relate to controlling resource consumption of a server and storage array. In embodiments, a request can be received by a server that is communicatively coupled to a storage array. Further, the services required to process the request can be identified. Additionally, services' activation can be controlled based on a mapping of request-related actions and initiated services.