Load Balancer Host Selection Using Response Time Anomaly Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing load balancing algorithms, such as the least-connections algorithm, can incorrectly select faulty servers in failure scenarios, leading to increased latency and errors due to the server's rapid response times and closed connections, causing further overload and performance issues.
Innovation Solution
Modifying the load balancer's host selection algorithm to incorporate statistical analysis of recent traffic data, including sampled response time averages and rate of change analysis, to detect anomalies and prevent routing requests to potentially faulty servers, and implementing remedial actions like disabling or restarting the host service.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the least-connections algorithm is used to distribute requests, then load distribution efficiency is improved, but reliability deteriorates when a faulty server responds rapidly with closed connections
Solution Approach 1:
The patent implements feedback mechanisms by monitoring server response characteristics and using this information to adjust load balancing decisions. The system tracks response times and connection patterns, feeding this data back into the host selection algorithm to identify and avoid faulty servers that exhibit anomalous rapid response behavior, thus resolving the contradiction between load distribution efficiency and request processing reliability.
Solution Approach 2:
The patent replaces the simple mechanical least-connections algorithm with a more sophisticated system that incorporates statistical analysis and anomaly detection. Instead of relying solely on connection counting, the system substitutes a comprehensive evaluation mechanism that considers response time patterns, rate of change metrics, and deviation from normal behavior to select healthy servers, thereby maintaining productivity while improving reliability.
2Reliability
If statistical analysis of recent traffic data is incorporated into host selection, then reliability is improved by detecting faulty servers, but device complexity increases
Solution Approach 1:
The patent applies partial action by implementing statistical analysis selectively rather than comprehensively. The system monitors key metrics such as response time averages and rate of change for suspicious servers, rather than analyzing all possible traffic data parameters. This approach achieves sufficient fault detection reliability while avoiding the excessive complexity that would result from complete traffic analysis, balancing reliability improvement with manageable algorithm complexity.
Data Source
AI summary
In order to protect against various load balancing failures, the host selection algorithm on the load balancer can be modified to take into account data available about the state of the entire service and each host server in the cluster. The state can include a number of metrics, including the sampled response time taken by the selected host service. The load balancer can use the state information in order to detect anomalies among the host services. For example, the load balancer can determine that the sampled response time of one host service has deviated by more than a standard deviation limit (or other predetermined threshold) from the sampled response times of the other host services in the cluster. If such an anomaly is detected, the load balancer can take various remedial actions, such as disabling the routing of incoming requests to the potentially faulty host service.