Dynamic Resource Node Deployment in Distributed Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing distributed computing systems face inefficiencies in resource deployment and utilization, leading to increased costs due to the static allocation of resources based on anticipated peak demands, resulting in underutilization during non-peak periods.
Innovation Solution
A resource management node that dynamically deploys and shuts down resource nodes based on real-time request loads and performance parameters, using a load balancer to distribute requests and adjust the number of active nodes within defined tolerance thresholds, optimizing resource allocation and utilization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If resources are statically allocated based on anticipated peak demands, then system reliability is improved, but resource utilization deteriorates during non-peak periods
Solution Approach 1:
The patent implements dynamic resource allocation by continuously monitoring system load metrics and automatically adjusting the number of active resource nodes based on current demand. This replaces the static allocation approach with a dynamic system that can adapt resource levels in real-time, maintaining reliability during peak periods while reducing resource consumption during non-peak periods.
Solution Approach 2:
The system employs feedback mechanisms by monitoring load metrics and using this information to control resource deployment. The load balancer continuously assesses system state and adjusts resource allocation accordingly, creating a closed-loop control system that responds to actual demand rather than relying on predetermined static allocation.
2Productivity
If the number of resource nodes is increased to handle peak demand, then system capacity is improved, but cost increases due to underutilization during non-peak periods
Solution Approach 1:
The system dynamically adjusts resource node count based on actual demand patterns, deploying additional nodes only when load metrics indicate the need for increased capacity. During non-peak periods, excess nodes are deactivated or shut down, preventing waste of computational resources and associated costs while maintaining adequate capacity during high-demand periods.
Solution Approach 2:
The patent changes the operational state of resource nodes by adjusting deployment parameters based on monitored load conditions. The system modifies the number of active nodes as a variable parameter rather than maintaining a fixed count, allowing capacity to scale with demand and reducing costs during low-utilization periods.
3Ease of operation
If load balancers are used to distribute requests, then resource balancing is improved, but efficiency of resource deployment deteriorates
Solution Approach 1:
The patent introduces a resource management node as an intermediary between the load balancer and resource nodes. This intermediary layer adds intelligence to the system by monitoring load metrics and controlling resource deployment based on actual system state, thereby improving deployment efficiency while the load balancer continues to handle request distribution for operational ease.
Data Source
AI summary
Some embodiments disclosed herein are directed to a resource management node that controls deployment of resource nodes in a distributed computing system. The resource management node includes a processor and a memory having computer readable program code that when executed by the processor causes the processor to perform operations that include deploying a group of resource nodes in a distributed computing system, and distributing requests received from the user electronic devices to the resource nodes of the group for processing. A resource deployment threshold is determined based on performance parameters of the resource nodes of the group, and a loading metric is determined based on the request. The operations include controlling further deployment of resource nodes to the group responsive to comparison of the loading metric to the resource deployment threshold. Related methods in a resource management node are disclosed.