Open Flow Network Topology Detection via KeepAlive Packets
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing open flow network faces challenges in maintaining and updating the detected topology, leading to increased load on the secure channel network, delayed detection of topology changes, and potential erroneous detection of non-occurring changes due to resource limitations in switches, which can cause system failures.
Innovation Solution
The introduction of a 'topology KeepAlive flow entry' and 'topology KeepAlive packet' mechanism, where switches maintain a circulation packet that circulates between them, allowing the controller to detect failures by monitoring the packet's arrival, reducing the load on the secure channel and enabling quicker detection of topology changes without relying on frequent scanning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If the controller frequently scans the network to detect topology changes, then the detection speed improves, but the load on the secure channel network increases
Solution Approach 1:
The controller sends topology discovery packets at regular intervals to detect topology changes. This periodic scanning approach balances detection speed with network load by not continuously monitoring but rather at scheduled times, reducing overall traffic while maintaining acceptable detection responsiveness.
Solution Approach 2:
Switches autonomously generate and send topology change notifications to the controller when they detect topology changes themselves, eliminating the need for the controller to continuously scan. This self-service mechanism reduces controller-initiated traffic while maintaining fast detection of actual changes.
2Measurement precision
If the controller increases scanning frequency to detect topology changes quickly, then the detection precision improves, but the switch resource consumption increases
Solution Approach 1:
Switches autonomously detect topology changes and send notifications to the controller, eliminating the need for frequent controller scanning. This reduces switch processing overhead while maintaining accurate detection of actual topology changes through event-driven notification.
Solution Approach 2:
The system implements a feedback mechanism where switches monitor their own topology state and provide feedback to the controller only when changes occur. This feedback-based approach ensures accurate detection without requiring continuous active scanning that would consume switch resources.
3Stability of the object's composition
If the controller reduces scanning frequency to lower network load, then the network stability improves, but the topology change detection delay increases
Solution Approach 1:
Switches autonomously detect and report topology changes to the controller, eliminating detection delay associated with periodic scanning intervals. This self-service approach maintains network stability by reducing overall traffic while achieving immediate detection of topology changes through event-driven notifications.
Solution Approach 2:
Switches continuously monitor their connectivity state in the background and are ready to immediately notify the controller when a topology change occurs. This preliminary monitoring preparation ensures fast detection without requiring active scanning at the moment of change, balancing stability and responsiveness.
4Measurement precision
If switches process more topology discovery packets to maintain accurate topology information, then the measurement precision improves, but the switch resource limitations cause erroneous detection
Solution Approach 1:
Switches autonomously generate topology change notifications based on their own state monitoring, reducing the number of processing packets required. This self-service approach maintains accurate topology information while minimizing processing load, preventing erroneous detections caused by resource exhaustion.
Solution Approach 2:
The system extracts only the essential topology change information from switches and sends it to the controller, rather than processing and analyzing all topology discovery packets. This extraction approach maintains detection accuracy while reducing processing requirements, preventing errors from resource limitations.
Data Source
AI summary
In an open flow network, a load of a switch on a secure channel network when a controller maintains and updates a physical topology of the switch in the situation that a network among the switches is in a high delay state. More specifically, the controller sets a flow entry having a rule and an action defined to uniformly control a packet as a flow to each of a plurality of switches. The controller sets to each switch, a circulation flow entry to be deleted when a circulation packet which is mutually transmitted and received among the switches gets not to arrive. Then, when receiving a notice that the circulation flow entry has been deleted, from each switch, the controller detects a failure among the switches.