Third-Party Network Computing Resource Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Enterprises face inefficiencies and higher costs when using reserved and on-demand computing resources from third-party networks, often reserving insufficient or excessive resources, leading to increased expenses due to fluctuating computational needs.
Innovation Solution
A remote network management platform analyzes past usage data to adjust the allocation of reserved and on-demand computing resources, calculating average hourly usage to determine optimal combinations of reserved and on-demand resources that meet demand while minimizing costs, ensuring no more than a threshold of on-demand resources are used.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If enterprises reserve computing resources in advance via reserved instances, then cost is reduced due to discounts, but resource allocation efficiency deteriorates when usage patterns change
Solution Approach 1:
The system dynamically adjusts the mix of reserved and on-demand computing resources based on monitored usage patterns. Instead of static allocation, the platform continuously learns from historical data and automatically reallocates resources to match actual enterprise needs, resolving the contradiction between cost savings from reserved instances and adaptability to changing usage patterns
Solution Approach 2:
The platform implements a feedback loop by monitoring actual computing resource usage, comparing it against reserved capacity, and automatically adjusting the allocation strategy. This closed-loop system uses usage data to refine future reservations, ensuring both cost optimization and adaptability to evolving enterprise requirements
2Adaptability or versatility
If enterprises reserve insufficient computing resources, then adaptability is improved, but cost increases due to higher on-demand pricing
Solution Approach 1:
The system performs preliminary analysis of usage patterns to determine optimal reservation levels before committing to reserved instances. By predicting future usage based on historical data, the platform pre-configures the right mix of reserved and on-demand resources, avoiding both over-reservation and under-reservation scenarios that lead to wasted spending
3Reliability
If enterprises reserve excessive computing resources, then reliability is improved, but cost increases due to over-provisioning
Solution Approach 1:
The system continuously adjusts the parameters of resource allocation based on monitored usage patterns. By changing the proportion of reserved versus on-demand resources dynamically, the platform maintains sufficient resource availability for reliability while eliminating excessive provisioning that would otherwise increase costs unnecessarily
Data Source
AI summary
A remote network management platform may include a database containing records relating to units of reserved and on-demand computing resources provided by a third-party network and a processor disposed within a computational instance. The processor may be configured to obtain utilization reports regarding the managed network from the third-party network and calculate, for each hour-of-day across one or more days of usage, respective hourly average units of utilization, by the managed network, of the reserved and on-demand computing resources. The processor may also calculate output values respectively associated with different combinations of the reserved and on-demand computing resources that jointly satisfy the hourly average units of utilization and select an allocation of the reserved computing resources that is within a threshold of a minimum output value of the output values. The processor may further change the number of allocated units to be the selected allocation of reserved computing resources.