Resource Manager for Dynamic Allocation in Cloud Hosts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing shared environments and cloud solutions face challenges with increasing energy consumption and complexity in managing resources, particularly in determining adequate resource allocation and monitoring applications' performance to meet Service Level Agreements (SLAs).
Innovation Solution
A method and apparatus that utilize a resource manager to determine and monitor resource component metrics, generate resource frames with time slots, and allocate resources based on these metrics to optimize performance and adherence to SLAs, allowing for better infrastructure management and energy efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If resource metrics are monitored and collected from multiple hosts to enable dynamic resource allocation, then resource utilization efficiency is improved, but system complexity and energy consumption increase
Solution Approach 1:
The patent segments resource monitoring into discrete resource frames with time slots for different resource types (CPU, memory, storage, network). Each host collects metrics independently and structures them into standardized frames, which the resource manager then processes separately for each resource type. This segmentation reduces system complexity by breaking down the monolithic monitoring task into manageable, modular components.
Solution Approach 2:
The patent introduces a resource manager as an intermediary between hosts and the resource allocation system. The resource manager collects standardized resource frames from multiple hosts, processes the metrics, and determines resource allocation decisions. This intermediary layer simplifies the overall system architecture by centralizing the complex logic of resource analysis and allocation, while hosts maintain simple, standardized metric collection and reporting functions.
2Reliability
If comprehensive resource metrics are collected from all hosts to ensure adequate SLA monitoring, then service level agreement compliance is improved, but energy consumption increases
Solution Approach 1:
The patent implements periodic resource frame collection with defined time slots for different resource types. Instead of continuous monitoring, the system collects metrics at regular intervals structured in resource frames. This periodic approach ensures SLA compliance by capturing resource usage patterns over time while reducing energy consumption compared to continuous monitoring, as hosts and the resource manager can enter low-power states between collection cycles.
Solution Approach 2:
The patent collects resource metrics for specific time slots corresponding to different resource types (CPU, memory, storage, network) rather than continuously monitoring all resources at all times. The resource manager determines allocation based on the collected frames, which may include more data than immediately necessary, allowing for historical analysis and trend detection while managing energy consumption through selective processing of metric data.
3Productivity
If resource allocation is determined based on collected metrics to optimize performance, then application performance is improved, but the complexity of resource management increases
Solution Approach 1:
The patent structures resource metrics into pre-defined resource frames with time slots for different resource types before analysis. Hosts collect and organize metrics in this standardized format in advance, and the resource manager receives pre-structured data that simplifies subsequent allocation decisions. This preliminary structuring of data reduces the complexity of resource management by eliminating the need for ad-hoc data organization and enabling systematic allocation algorithms.
Solution Approach 2:
The patent transforms raw resource metrics into standardized resource frames with specific parameters for different resource types (CPU usage, memory consumption, storage I/O, network traffic). The resource manager processes these parameterized frames to determine allocation, changing the form of the data from heterogeneous host-specific formats to a unified parameter structure. This parameter transformation simplifies resource management complexity by enabling consistent analysis and comparison across different hosts and resource types.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Method, resource manager, and computer program performed by a resource manager (110) in a communications network (50) connected to at least two hosts (120) for enabling of resource component allocation related to the hosts (120), comprising determining a metric of a resource component (130) for performance monitoring of an application at the hosts (120),transmitting an instruction to measure the metric to respective host (120),receiving a respective resource frame (170) of the application(155) from respective host (120),determining resource allocation for the application (155) based on the received resource frames (170).