Cloud Resource Management via Dynamic Capacity Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cloud computing and communication resource providers face challenges in managing variable demand patterns, leading to inefficient capacity utilization and increased costs due to congestion, which violates service level agreements and requires costly capacity upgrades.
Innovation Solution
A data management process that predicts resource requirements by analyzing historical user reservations and actual usage patterns to determine projected capacity needs, allowing for dynamic resource allocation and configuration to meet demand, while also considering contingency surpluses and unreserved requests.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If capacity is increased to meet peak demand, then service level agreements are met, but costs increase disproportionately
Solution Approach 1:
The patent implements dynamic resource allocation where capacity is adjusted in real-time based on actual demand patterns. The system monitors usage continuously and reallocates resources dynamically, transitioning from static over-provisioning to adaptive provisioning that matches actual needs, thereby reducing unnecessary capacity costs while maintaining SLA compliance during peak periods.
Solution Approach 2:
The system changes key parameters such as capacity allocation, pricing rates, and resource provisioning levels based on demand conditions. By adjusting these parameters dynamically rather than maintaining fixed values, the system optimizes the balance between reliability and cost, allocating sufficient capacity during peaks and reducing it during low-demand periods.
2Reliability
If resources are reserved in advance, then performance guarantees are met, but resource utilization efficiency decreases
Solution Approach 1:
The system performs preliminary actions by reserving capacity in advance based on predicted demand patterns and SLA requirements. However, it optimizes this preliminary reservation by using historical data and forecasting to allocate only the necessary amount of reserved capacity, avoiding excessive reservations that would reduce utilization efficiency while still ensuring performance guarantees.
Solution Approach 2:
The system implements feedback mechanisms that monitor actual resource usage against reservations and SLA commitments. This feedback loop allows the system to adjust future reservation levels based on actual performance and utilization patterns, optimizing the balance between guaranteeing performance and maintaining efficient resource utilization over time.
3Reliability
If demand is managed conservatively, then service quality is maintained, but capacity costs increase
Solution Approach 1:
The patent replaces conservative static capacity planning with dynamic capacity management that adapts to actual demand. The system continuously monitors service quality metrics and adjusts capacity allocation in real-time, maintaining high service quality during critical periods while reducing capacity investment during periods of lower demand, thereby eliminating the need for consistently over-provisioning.
Data Source
AI summary
A data management process determines, from user-implemented provisional reservations (400) for data processing resources, a projected total capacity requirement for each said data processing resource, by maintaining a record (9, 90, 91) recording previous such reservations made by each user and comparing each reservations with records (87, 88, 89) of the actual resources used, to provide an estimate of resources required to meet the projected capacity requirement, and to provide data for a demand management processor (2), which control associated configurable data processing equipment (1) to provide the resources required to meet the estimated capacity required. The process takes account of over- and under-ordering of capacity by comparing each reservation (400) with the use actually made (600), and includes a record (10) of ad-hoc (unreserved) usage.