Routing Destination Evaluation Using Time-Series Metrics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing load distribution techniques, particularly static schemes, face efficiency losses when changes occur in server states, while dynamic schemes require continuous monitoring and load balancer usage for real-time evaluation of routing destinations.
Innovation Solution
A routing destination evaluation apparatus that acquires time-series data of various metrics associated with potential routing destinations, evaluates them based on these data, and generates evaluation data, allowing for dynamic evaluation and prioritization of candidates, thereby adjusting the influence of metric values on the evaluation process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a static distribution scheme is used, then the load distribution can be realized with a simply policy, but the efficiency is lowered when changes occur in the states of routing destination candidates
Solution Approach 1:
The patent implements dynamic evaluation of routing destination candidates by continuously acquiring time-series metric data and recalculating evaluation values based on current server states. This allows the system to adapt to changing conditions automatically, resolving the contradiction by making the distribution scheme dynamic rather than static, thereby maintaining efficiency without sacrificing operational simplicity through automated decision-making
2Productivity
If a dynamic distribution scheme is used, then the load distributing efficiency is not easily lowered even if there is a change in the state of the routing destination candidates, but a load balancer is required to monitor the states of the routing destination candidates in real time and select a routing destination
Solution Approach 1:
The patent extracts the core evaluation function from a traditional load balancer architecture by implementing a dedicated evaluation apparatus that separately monitors metric data and calculates evaluation values. This separation allows the routing control apparatus to make decisions based on evaluation data without requiring complex real-time monitoring infrastructure, thereby maintaining efficiency while reducing device complexity through functional decomposition
3Measurement precision
If multiple types of metrics are considered for evaluation, then the evaluation comprehensiveness is improved, but the complexity of evaluation process increases
Solution Approach 1:
The patent handles multiple metric types by normalizing them to a common evaluation scale and applying weighted sums based on priority levels. This parameter transformation approach allows comprehensive evaluation of diverse metrics (server load, response time, error rates, etc.) while maintaining process simplicity through standardized calculation methods, resolving the contradiction by changing the parameter representation rather than the evaluation structure
Data Source
AI summary
According to one embodiment, a routing destination evaluation apparatus includes a metrics acquisition unit and a routing destination evaluation unit. The metrics acquisition unit acquires time-series data of types of metrics associated with a candidate. The candidate is one of candidates of a routing destination to which the routing control apparatus routes a request including the first destination data. The routing destination evaluation unit evaluates the candidates based on the time-series data, and generates routing destination evaluation data.