Virtual Machine Placement in Cloud Data Centers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Network service providers face challenges in efficiently managing and allocating finite computing resources in dynamic and complex cloud-computing environments, where resources like processing power and memory are confined to individual physical machines, while data storage is pooled across multiple machines.
Innovation Solution
The method involves real-time adaptive placement of virtual machines within a cloud-based network, where virtual machines are routed to a target data center based on a data center index calculation and then assigned to a physical machine based on a configuration index, minimizing maximum average resource utilization across data centers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If virtual machines are placed in a cloud-based network with multiple data centers, then resource sharing and flexibility are improved, but the complexity of managing and allocating finite computing resources increases
Solution Approach 1:
The patent introduces a centralized controller as an intermediary that manages virtual machine placement across multiple data centers. This controller receives placement requests, calculates optimal destinations using index formulas that consider resource utilization metrics, and coordinates the placement process. By centralizing control logic, the system improves resource sharing flexibility while managing complexity through a single coordination point rather than distributed decision-making.
Solution Approach 2:
The patent employs dynamic parameter changes by using index formulas that calculate data center destinations based on real-time resource utilization metrics. The system monitors parameters such as average resource utilization, queue lengths, and configuration usage fractions, then uses these changing parameters to dynamically determine optimal virtual machine placement. This allows the system to adapt to varying load conditions and maintain efficient resource allocation across the distributed infrastructure.
2Reliability
If resources are confined to individual physical machines for processing and memory, then resource allocation control is improved, but resource utilization efficiency deteriorates
Solution Approach 1:
The patent implements universality by enabling physical machines to host multiple virtual machines with different configurations and workloads. Each physical machine serves multiple functions by running diverse virtualized workloads, thereby improving resource utilization efficiency. The centralized controller manages this multi-functionality by allocating virtual machines to physical machines based on resource availability and utilization metrics, maintaining control while maximizing productivity.
Solution Approach 2:
The system applies dynamics by enabling virtual machines to be migrated between physical machines based on changing resource utilization conditions. The centralized controller continuously monitors resource metrics and dynamically repositions virtual machines to optimize utilization. This dynamic approach allows the system to maintain reliable resource allocation control while adapting to varying workload demands, thereby improving overall resource utilization efficiency.
3Adaptability or versatility
If data storage is provided as a pooled service across multiple physical machines, then resource sharing is improved, but the difficulty of managing heterogeneous resources increases
Solution Approach 1:
The centralized controller acts as an intermediary that manages pooled storage resources across multiple physical machines. It receives storage-related placement decisions, calculates optimal destinations using index formulas that consider resource utilization, and coordinates storage resource allocation. This intermediary approach enables efficient sharing of pooled storage resources while managing the complexity of heterogeneous resources through centralized coordination and standardized management protocols.
Data Source
AI summary
Methods and apparatuses for real-time adaptive placement of a virtual machine are provided. In an embodiment, a virtual machine is received at a routing component, the routing component having a processor in communication with a memory. By the processor in communication with the memory, a target data center is determined from a plurality of data centers based on a data center index, and the virtual machine is routed to the target data center. A physical machine is chosen within the target data center for placing the virtual machine.