Datacenter Resource Manager for Provisioning Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cloud computing users face inefficiencies in determining the required resources for their applications, leading to potential under or over-provisioning due to uncertainties in performance caused by shared resources and activities of other users in multi-resource, multi-tenant datacenters.
Innovation Solution
A resource manager element that maps user requests to candidate resource combinations using benchmarking and analytical models, providing data on these combinations to users to help them make informed decisions, and automatically provisioning resources based on selected options.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If users request more resources to ensure performance requirements are met, then reliability is improved, but resource utilization efficiency deteriorates due to over-provisioning
Solution Approach 1:
The system performs preliminary actions by proactively predicting future performance requirements and resource needs before users actually request resources. The performance prediction module analyzes historical data and current system state to forecast when performance issues may arise, allowing the resource allocation module to pre-allocate or pre-release resources, thereby avoiding both over-provisioning and under-provisioning.
Solution Approach 2:
The system implements continuous feedback loops where the performance prediction module constantly monitors actual system performance against predicted performance. This feedback information is used to refine predictions and adjust resource allocation dynamically. The resource allocation module receives feedback on resource utilization and performance outcomes, enabling it to optimize future allocation decisions and improve overall resource efficiency.
2Loss of energy
If users request fewer resources to improve efficiency, then resource utilization efficiency is improved, but reliability deteriorates due to under-provisioning
Solution Approach 1:
The system applies dynamics by making resource allocation flexible and adaptive rather than static. The resource allocation module continuously adjusts resource provisioning based on real-time performance predictions and actual system state. This dynamic approach allows the system to scale resources up or down as needed, ensuring reliability is maintained while optimizing resource utilization efficiency through just-in-time resource allocation.
3Device complexity
If manual resource provisioning is used, then device complexity is reduced, but productivity deteriorates due to time-consuming resource determination
Solution Approach 1:
The system implements self-service by enabling automatic resource provisioning without requiring manual user intervention. The performance prediction module autonomously analyzes system state and predicts resource needs, while the resource allocation module automatically translates these predictions into resource allocation decisions. This self-service mechanism significantly improves productivity by eliminating time-consuming manual resource determination processes while maintaining manageable system complexity through automated decision-making.
4Productivity
If automated resource provisioning is implemented, then productivity is improved, but device complexity increases due to prediction and allocation systems
Solution Approach 1:
The system applies segmentation by dividing the complex resource provisioning task into distinct functional modules: a performance prediction module that handles analysis and forecasting, and a resource allocation module that handles decision-making and provisioning. This segmentation allows each module to specialize in specific functions, reducing the complexity burden on any single component while maintaining high overall productivity through coordinated automated operation.
Data Source
AI summary
Methods of predicting datacenter performance to improve provisioning are described. In an embodiment, a resource manager element receives a request from a tenant which describes an application that the tenant wants executed by a multi-resource, multi-tenant datacenter. The request that has been received is mapped to a set of different candidate resource combinations within the datacenter, where each candidate resource combination can be used to execute the application in a manner which satisfies a high level constraint specified within the request. This mapping may, for example, be performed using a combination of benchmarking and an analytical model. In some examples, each resource combination may comprise a number of virtual machines and a bandwidth between those machines. Data relating to at least a subset (and in some examples, two or more) of the candidate resource combinations is then presented to the tenant.