Cloud Resource Configuration via Bidding Instance Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current computing resource configuration methods for enterprise cloudification, such as on-demand and reserved instances, result in high Total Cost of Ownership (TCO) due to inefficient resource utilization and pricing models, particularly when dealing with big data processing and rapid data generation.
Innovation Solution
A computing resource configuration method and device that utilizes a bidding instance model, where users configure the proportion of bidding instances to on-demand instances, dynamically determines required resources, and applies for them from a cloud provider, incorporating a rebalancing mechanism to optimize resource allocation and reduce TCO by leveraging market price fluctuations and idle resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If on-demand instances are used for computing tasks, then computing resources can be obtained quickly and flexibly, but the Total Cost of Ownership (TCO) increases significantly
Solution Approach 1:
The patent implements dynamic instance type selection that adjusts between bidding instances, reserved instances, and on-demand instances based on real-time workload characteristics, resource availability, and cost considerations. The system continuously monitors computing task requirements and automatically selects the most cost-effective instance type, transforming the static resource allocation into a dynamic optimization process that reduces TCO while maintaining flexibility.
2Loss of energy
If reserved instances are used for computing tasks, then cost is reduced compared to on-demand instances, but resource allocation flexibility decreases
Solution Approach 1:
The patent segments the computing resource pool into multiple instance type categories (bidding instances, reserved instances, on-demand instances) and assigns different tasks to appropriate segments based on their characteristics. Workload tasks are classified and routed to specific instance types: stable long-running tasks use reserved instances for cost efficiency, while variable or urgent tasks use bidding or on-demand instances for flexibility, achieving both cost reduction and adaptability through segmented resource management.
3Loss of energy
If bidding instances are used to reduce TCO, then cost efficiency improves, but resource availability may be affected by market price fluctuations and downtime events
Solution Approach 1:
The patent implements a rebalancing mechanism that monitors resource availability and preemptively adjusts instance allocations before downtime or resource recovery events impact task execution. When market price fluctuations or cloud provider downtime events are detected, the system proactively reallocates tasks from affected bidding instances to alternative instance types, cushioning against potential service disruptions and maintaining task completion reliability despite resource availability variations.
4Productivity
If computing resources are dynamically allocated based on task requirements, then resource utilization efficiency improves, but system complexity increases
Solution Approach 1:
The patent implements self-service automation where the system autonomously performs instance type selection, task routing, and resource rebalancing without requiring manual intervention. The automated orchestration engine continuously monitors task requirements, evaluates available resources across different instance types, and automatically allocates computing resources based on predefined optimization criteria, achieving high resource utilization efficiency while managing system complexity through automation rather than manual processes.
Data Source
AI summary
An embodiment of the present invention discloses a computing resource configuration method and device for enterprise cloudification. The method comprises the following steps: obtaining parameters configured by a user through a configuration page, wherein the parameters include a proportion of a bidding instance to an on-demand instance; determining a total amount of computing resources required by a computing task after the computing task is obtained; determining the computing resources corresponding to the bidding instance and the computing resources corresponding to the on-demand instance based on existing computing resources, the total amount of the computing resources and the proportion; and applying for the computing resources from a cloud provider based on the computing resources corresponding to the bidding instance and the computing resources corresponding to the on-demand instance. Big data is computed in a bidding instance mode so that the total cost of the computing resources is greatly reduced. Moreover, the technical problem of high total cost of the computing resources caused by performing data computing in a mode of depending on a reserved instance and/or the on-demand instance is solved.