VIP State Notification for Global Server Load Balancing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Global server load balancing systems experience prolonged failure detection times due to periodic polling methods for virtual IP address status, leading to potential network outages during intervals between keepalive checks.
Innovation Solution
Implementing a VIP address state notification mechanism that allows real-time notification from an application control engine to a global site selector, reducing the need for periodic keepalive checks and enabling immediate state change reflection, thereby minimizing network outages.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If periodic polling methods are used to check VIP address status, then the system can detect failures, but the failure detection time is prolonged due to intervals between keepalive checks
Solution Approach 1:
Instead of the global site selector periodically polling the application control engine for VIP status (active probing), the system inverts the approach by having the application control engine actively notify the global site selector of VIP state changes (passive notification). This inversion eliminates the need for continuous polling and enables immediate detection of VIP failures as they occur.
Solution Approach 2:
The application control engine performs preliminary monitoring of VIP address states and is prepared to immediately notify the global site selector when a state change occurs. This preliminary action ensures that failure detection happens at the moment of change rather than waiting for the next scheduled poll, reducing detection time while maintaining system reliability.
2Reliability
If periodic keepalive checks are implemented to monitor VIP address status, then failure detection is possible, but the processing load on keepalive engines increases
Solution Approach 1:
The system extracts the continuous polling function from the global site selector's keepalive engine and transfers it to the application control engine, which monitors VIP states and sends notifications only when changes occur. This extraction eliminates the need for ongoing periodic checks by the keepalive engine, significantly reducing its processing load while maintaining reliable VIP status monitoring through event-driven notifications.
Solution Approach 2:
The application control engine takes responsibility for monitoring VIP address states and self-manages the notification process to the global site selector. This self-service approach allows the system to monitor VIP status continuously without requiring external polling, reducing the processing burden on the keepalive engine while ensuring reliable failure detection.
Data Source
AI summary
Methods and systems are provided for global server load balancing in a communication network using a global site selector and an application control engine in communication with the global site selector, in which the application control engine notifies the global site selector in response to a state change for a virtual internet protocol (VIP) address. Keepalive methods generally rely on periodically polling VIP addresses to check whether they are in an ONLINE or OFFLINE state. A minimum interval between two keepalives can be as long as 40 seconds, during which, without a mechanism to reflect state change to the load balancer in real time, if a VIP address changes state (e.g., VIP address failure), a customer can experience network outage during the time interval. Embodiments provide a mechanism to reflect, in real time, the current VIP address state to a global server load balancer to reduce failure detection time.