Virtualized Data Center Resource Provisioning for Energy and Priority Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In virtualized data centers, the preemption of best-effort requests for advance-reservation requests can lead to starvation of low-priority requests, necessitating a method to prevent such starvation while optimizing energy consumption.
Innovation Solution
A system and method that computes a violation risk factor for low-priority requests and evaluates the utilization of activated resources, using a fuzzy engine to provision resources by preempting, activating new resources, or consolidating virtual machines based on predefined rules to ensure fair and energy-efficient allocation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If best-effort requests are preempted for advance-reservation requests, then resource availability for high-priority requests is improved, but low-priority requests experience starvation
Solution Approach 1:
The system continuously monitors the waiting time of low-priority requests and uses this feedback to dynamically adjust preemption decisions. When the waiting time of low-priority requests approaches their threshold, the system reduces or pauses preemption actions, thereby preventing starvation while still allowing necessary resource reallocation for high-priority requests.
Solution Approach 2:
The preemption policy is made dynamic rather than static. The system adjusts the preemption behavior based on real-time conditions such as resource availability, priority levels, and waiting time thresholds. This dynamic adjustment allows the system to balance between serving high-priority requests and preventing low-priority request starvation.
2Productivity
If preemption is used to serve high-priority requests, then resource allocation efficiency is improved, but energy consumption increases due to frequent resource activation
Solution Approach 1:
The system changes the parameters of resource provisioning by introducing a violation risk factor that combines waiting time and resource utilization considerations. This parameter modification allows the system to make more informed decisions about when to preempt and when to activate new resources, balancing allocation efficiency with energy consumption.
Solution Approach 2:
The system performs preliminary evaluation of resource utilization factors and violation risk factors before executing preemption or activation decisions. This preliminary action allows the system to predict the consequences of different actions and choose those that minimize energy consumption while maintaining allocation efficiency.
3Reliability
If resource activation is increased to meet high-priority requests, then service level agreements are improved, but energy waste increases
Solution Approach 1:
The system modifies the resource provisioning parameters by incorporating both violation risk factor and resource utilization factor. This dual-parameter approach enables the system to activate resources only when necessary, ensuring SLA compliance while minimizing energy waste through more precise resource allocation decisions.
Data Source
AI summary
A method and system provisions a plurality of resources of a data center. A violation risk factor for a set of low priority requests can be computed. A utilization factor of a set of activated resources of the data center shall be evaluated. According to a predefined rule base, one or more of the plurality of resources, shall be provisioned for a received high priority request, whereby the predefined rule base defines performing one or more of; a) preempting a set of virtual machines utilizing a subset of the set of activated resources, whereby the set of virtual machines is associated with the set of low priority requests; b) activating a new set of resources; and c) consolidating a plurality of virtual machines, based on the computed violation risk factor and the evaluated utilization factor.