AI-Driven Elastic Network Scaling for SLA-Aware Energy Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing networks face inefficiencies in energy consumption due to overprovisioning, leading to idle resources and increased energy use, while under-provisioning risks SLA violations and poor user experience, with no mechanisms to balance resource consumption with SLA/QoE satisfaction.
Innovation Solution
Implementing an AI-driven elastic network that dynamically adapts its architecture to meet traffic demand, using digital twins and machine learning to optimize energy use while maintaining performance constraints.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If network capacity is significantly expanded through overprovisioning to meet future demand, then network design and operation are simplified, but energy consumption increases due to idle resources
Solution Approach 1:
The patent implements dynamic network resource allocation where network elements can be activated or deactivated based on real-time traffic demand. The system transitions from static overprovisioned networks to dynamic networks that adapt their capacity, allowing resources to be turned off during low-demand periods while maintaining simplicity through automated control mechanisms.
Solution Approach 2:
The system changes operational parameters of network elements by adjusting their active/inactive states based on traffic patterns. Through parameter changes in resource allocation and element activation, the network maintains simplicity of operation while reducing energy consumption by only activating resources when needed.
2Use of energy by moving object
If network capacity is reduced to decrease energy consumption, then energy efficiency improves, but SLA violations occur and network performance deteriorates
Solution Approach 1:
The system performs preliminary actions by pre-provisioning resources and using predictive algorithms to anticipate future traffic demands. This allows the network to maintain adequate capacity during peak periods without permanently overprovisioning, thereby reducing energy consumption while ensuring SLA compliance through proactive resource allocation.
Solution Approach 2:
The patent implements feedback mechanisms where network performance metrics and traffic patterns are continuously monitored and fed back to the control system. This feedback loop enables the system to adjust resource allocation dynamically, ensuring that capacity is maintained only when needed to satisfy SLAs while minimizing energy consumption during periods when full capacity is not required.
3Reliability
If network resources are allocated to handle peak demand, then SLA requirements are met during high traffic periods, but energy consumption increases during low-demand periods
Solution Approach 1:
The system dynamically adjusts network resource allocation based on real-time and historical traffic patterns. During peak demand periods, resources are activated to meet SLA requirements, while during low-demand periods, resources are deactivated or put into sleep modes, thereby reducing energy consumption without compromising SLA satisfaction when needed.
Solution Approach 2:
The system uses predictive analytics and historical data to perform preliminary actions by pre-positioning resources or pre-warming network elements before peak demand periods occur. This allows the network to meet SLA requirements during peaks without maintaining full capacity continuously, reducing energy consumption during low-demand periods while ensuring adequate resources are ready when needed.
Data Source
AI summary
In one implementation, a device maintains a digital twin of a computer network. The device determines, based on the digital twin, an action to reduce energy consumption by the computer network. The device validates the action using the digital to ensure that performance of the action in the computer network will result in the computer network still satisfying one or more performance constraints. The device causes performance of the action in the computer network, when the action is deemed valid.