HCI Storage Management via Usable Space Calculation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In hyperconverged infrastructure (HCI) clusters, existing resource management systems struggle to accurately compute actual usable storage space and free space, leading to potential over-utilization or under-utilization, which can result in performance issues due to misleading raw storage information and neglect of user policy settings and compression ratios.
Innovation Solution
A management system that retrieves user policy settings from a database to compute actual usable storage and free space by adjusting raw storage information with factors like replication, deduplication, and compression ratios, and dynamically scales resources by cloning or terminating hosts based on revenue and expense ratios to maintain optimal utilization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If raw storage information is used without adjustment, then storage capacity appears sufficient, but actual usable storage is miscalculated leading to over-utilization
Solution Approach 1:
The patent introduces an intermediary calculation layer that adjusts raw storage information by applying policy factors (replication, deduplication, compression, maintenance reserves) to derive actual usable storage. This intermediary process mediates between the simple raw storage report and the complex reality of usable storage, resolving the contradiction by adding computational steps without requiring fundamental system redesign.
Solution Approach 2:
The system implements feedback by continuously monitoring storage utilization against calculated thresholds and triggering scaling actions (cloning or terminating hosts) when thresholds are breached. This feedback loop ensures storage capacity is dynamically adjusted based on actual usable storage calculations, preventing over-utilization while maintaining system simplicity through automated responses.
2Quantity of substance
If storage resources are scaled out by adding hosts, then storage capacity increases, but resource management analysis demand increases
Solution Approach 1:
The system enables self-service by automatically calculating storage needs, determining optimal host cloning or termination decisions, and executing scaling operations without manual intervention. The resource management system serves itself by monitoring its own storage utilization and autonomously adjusting capacity through host lifecycle management, reducing the need for external automation while handling increased storage capacity.
Solution Approach 2:
The patent applies preliminary action by pre-calculating storage thresholds and scaling triggers before utilization problems occur. The system establishes utilization thresholds based on policy factors in advance, and when thresholds are approached, it proactively clones or terminates hosts to maintain optimal utilization, preventing rather than reacting to storage issues.
3Quantity of substance
If hosts are cloned to provide additional storage, then storage capacity increases, but infrastructure cost increases
Solution Approach 1:
The system changes parameters by dynamically adjusting storage utilization thresholds based on policy factors (replication factor, deduplication ratio, compression ratio, maintenance reserves). By modifying these parameters, the system optimizes the balance between storage capacity and infrastructure cost, cloning hosts only when necessary based on calculated actual usable storage rather than raw capacity, thereby reducing unnecessary infrastructure expenditure.
Solution Approach 2:
The patent implements dynamics by making the storage scaling system adaptive and responsive to changing conditions. The resource management system continuously monitors utilization, recalculates thresholds as policies change, and dynamically adjusts host capacity through cloning or termination. This dynamic approach ensures storage capacity scales precisely when needed, avoiding both over-provisioning (wasted cost) and under-provisioning (performance degradation).
Data Source
AI summary
Various approaches for managing computational resources in a hyperconverged infrastructure (HCI) cluster include identifying the hosts associated with the HCI cluster for providing one or more computational resources thereto; for each of the hosts, determining a revenue and/or an expense for allocating the computational resource(s) to the HCI cluster; and determining whether to clone, suspend or terminate each host in the HCI cluster based at least in part on the associated revenue and/or expense.