Container Cost Attribution via Direct and Indirect Resource Metrics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The challenge lies in accurately determining and attributing the cost of containers running in data centers, particularly due to their short-lived nature and volatile association with host servers, which complicates cost measurement and distribution across different computing environments.
Innovation Solution
A method and system are developed to calculate the total container cost by combining direct and indirect costs, where direct costs are based on resource utilization and indirect costs account for host overhead, enabling accurate attribution to applications deployed within these containers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If containers are used to run applications in data centers, then application deployment flexibility and resource efficiency are improved, but cost measurement and attribution become difficult due to short-lived nature and volatile association with host servers
Solution Approach 1:
The system implements continuous monitoring of container resource utilization and host resource consumption, collecting data throughout the container lifecycle. This feedback mechanism tracks CPU, memory, storage, and network usage, enabling accurate cost calculation despite the transient nature of containers. The monitoring data feeds into cost attribution algorithms that distribute host costs to containers based on actual resource consumption patterns.
Solution Approach 2:
The patent introduces an intermediary cost attribution system that acts as a mediator between physical host resources and containerized applications. This intermediary layer collects resource utilization metrics from hosts, calculates proportional costs based on container usage, and attributes costs to applications. The intermediary resolves the measurement challenge by bridging the gap between volatile container associations and stable cost accounting requirements.
2Adaptability or versatility
If containers are moved between hosts for development and staging, then application portability is improved, but tracking utilization and attributing costs across multiple hosts becomes complex
Solution Approach 1:
The system implements a universal cost tracking mechanism that functions consistently across multiple hosts and container lifecycles. The cost attribution system maintains a unified view of container resource consumption regardless of which host the container is running on. This universal approach allows containers to be moved between development, staging, and production environments while maintaining continuous cost tracking and accurate attribution to the underlying applications.
Solution Approach 2:
The patent employs preliminary monitoring and data collection mechanisms that track container resource utilization from the moment a container is created or moved to a host. By collecting utilization data continuously and maintaining running totals, the system prepares cost attribution information in advance, eliminating the need for complex retrospective calculations when containers migrate between hosts or when cost reporting is required.
3Measurement precision
If monitor containers are added to track container utilization, then measurement accuracy is improved, but system resource overhead and total cost increase
Solution Approach 1:
The system merges the monitoring function into the existing container infrastructure by utilizing the host operating system's built-in resource tracking capabilities and virtualization management tools. Instead of adding separate monitor containers that would consume additional resources, the patent leverages data already collected by the host system for resource management and scheduling purposes. This merging approach enables accurate utilization measurement while minimizing additional resource overhead.
Data Source
AI summary
Methods and system to calculate a total container cost of each container of a data center are described. Methods compute a direct container cost of each container running in the data center based on container resource utilization and a resource utilization rate of each resource of a host. Methods also compute an indirect container cost of each container is calculated as a sum of indirect cost of each resource of the host used by the container and each resource of the host used by a monitor container that monitors each container utilization of host resources. For each container, a total container cost of the container is calculated as the sum of the direct container cost and the indirect container cost of the container. Total container costs are used to determine the cost of applications miming on the containers.