Network Flow Management via Aggregated Switch Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional Software Defined Networking (SDN) techniques face limitations in managing network flows across the network, failing to easily detect and address operational environment conditions, bandwidth requirements, positional configurations, and transient/sustained flow microbursts, among others.
Innovation Solution
An information handling system that collects first-level flow information from switch devices, filters and structures it to generate second-level flow information, analyzes it based on current flow operation policies, and determines policy changes to distribute updated policies to switch devices for optimized flow operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If conventional SDN techniques utilize individual switch-based and port-based data collection, then switch-level management and flow analysis are achieved, but network-wide flow management capability is limited
Solution Approach 1:
The patent introduces a flow collector as an intermediary component that aggregates flow information from multiple switches and presents a unified view to the controller. This mediator layer consolidates data from individual switch-based collection points into network-wide flow information, enabling comprehensive flow management without requiring each switch to independently perform complex analysis tasks.
Solution Approach 2:
The patent transitions from single-switch, single-port dimensional flow collection to a multi-dimensional aggregation model where flow information is collected across multiple switches, ports, and time periods. The flow collector aggregates data along multiple dimensions (switch identity, port identity, time windows) to create comprehensive network-wide flow views that enable holistic flow management.
2Speed
If real-time flow information collection is implemented across the network, then flow management responsiveness is improved, but system resource consumption increases
Solution Approach 1:
The patent implements periodic flow information collection where the flow collector gathers flow data at regular time intervals rather than continuously. Flow information is collected for specific time windows and aggregated periodically, providing timely flow management capabilities while reducing the overall data collection burden and resource consumption compared to continuous monitoring.
Solution Approach 2:
The patent collects flow information at a granular level (individual switch and port levels) that is more detailed than strictly necessary for network-wide management, then aggregates this excessive detail into summarized network-wide views. This partial collection approach ensures all relevant flow data is captured at the source while the aggregation process filters out redundant information for higher-level management.
3Loss of information
If comprehensive flow information is collected from all switches, then network-wide flow visibility is achieved, but data processing complexity increases
Solution Approach 1:
The patent segments the flow information collection and processing task into distinct hierarchical levels: individual switch-level flow data collection, flow collector-level aggregation across multiple switches, and controller-level network-wide flow management. This segmentation allows each component to handle only its portion of the data processing complexity while maintaining comprehensive network-wide flow visibility through the hierarchical aggregation structure.
Data Source
AI summary
A network flow management system includes controllers that are each coupled to a subset of switch devices, and a flow management server system that is coupled to each of the controllers. The flow management server system collects first-level flow information for the switch devices from the controllers, and filters and structures the first-level flow information to generate second-level flow information. The flow management server system then analyzes the second-level flow information based on current flow operation policies that are configured to cause the switch devices to perform first flow operations and, in response, determines flow operation policy changes. The flow management server system then distributes updated flow operation policies that includes the flow operation policy changes to each of the controllers, where the updated flow operation policies cause the switch devices to perform second flow operation that are different than the first flow operations.