Workload Distribution in Slice-Based Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current network slicing technologies lack the ability to optimally place virtual network functions (VNFs) in proximity to network workloads, leading to suboptimal performance due to limited visibility and access to resources across slice-managing components, resulting in inefficient allocation of virtual and physical resources.
Innovation Solution
Implement a system that uses inter-cloud resource federation to enable region discovery and workload distribution across multiple sites, allowing VIMs to monitor performance thresholds and instantiate VNFs in optimal locations based on workload proximity and resource availability, ensuring compliance with SLA metrics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If network slicing is implemented to allow multiple logical networks on shared physical infrastructure, then network resource utilization and connectivity are improved, but the ability to optimally place VNFs in proximity to workloads deteriorates due to limited visibility across slice-managing components
Solution Approach 1:
The patent introduces a workload placement optimization system that acts as an intermediary between multiple slice-managing components. This system receives workload placement requests, queries multiple slice-managing components for available resources and performance metrics, analyzes the responses to determine optimal placement locations, and coordinates the placement across the distributed architecture. The intermediary enables centralized optimization while maintaining the distributed slice-managing structure, resolving the contradiction between resource utilization and latency.
2Speed
If VNFs are placed closer to workload sources to reduce latency, then network performance is improved, but the complexity of managing resource allocation across distributed slice-managing components increases
Solution Approach 1:
The patent segments the workload placement management function into distinct modular components: a placement optimization system that handles high-level decision-making, individual slice-managing components that provide local resource information, and coordination mechanisms that facilitate communication. Each component has a specific, simplified responsibility, which reduces the complexity burden on any single component while enabling the overall system to achieve optimal VNF placement for improved data transmission speed.
3Productivity
If centralized management is used to optimize VNF placement, then workload distribution is improved, but the scalability and adaptability to different slice requirements deteriorate
Solution Approach 1:
The patent implements local quality by allowing each slice-managing component to maintain slice-specific configurations, policies, and resource characteristics while participating in the centralized workload placement optimization. The placement optimization system queries these local components for slice-specific information and incorporates it into placement decisions. This enables centralized coordination for efficient workload distribution while preserving local adaptability to different slice requirements such as IoT, mobile broadband, and vehicular communications.
4Reliability
If multiple slice-managing components operate independently to manage different logical networks, then network isolation and specialized management are improved, but visibility and access to resources across components deteriorates
Solution Approach 1:
The patent implements a feedback mechanism where the workload placement optimization system queries multiple slice-managing components for resource availability, performance metrics, and slice requirements, receives their responses, analyzes the combined information, and uses this feedback to make informed placement decisions. The slice-managing components also receive feedback about placement decisions and can adjust their local resource allocation accordingly. This feedback loop maintains network isolation while improving resource visibility across components.
Data Source
AI summary
A system incorporated in a slice-based network can implement a first virtual infrastructure manager (“VIM”) at a first region. The first VIM can be associated with a first internet protocol (“IP”) prefix range, and configured to receive a second IP prefix range associated with a second region having a second VIM. For compliance with requirements from a software license agreement (“SLA”), the first VIM can monitor a performance of a first virtual network function (“VNF”) of a network slice. In the event of a performance threshold violation, the first VIM can map portions of a workload associated with the violated threshold to the first region and the second region based on respective workload flow data associated with each of the first and second IP prefix ranges. The first VIM can instantiate a second VNF in the region having a workload portion that corresponds to a higher network resource consumption.