Dynamic Server Load Balancing via Age-Based Bias
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In data centers, 'focus firing' occurs when multiple load balancers simultaneously route workloads to newly added servers, leading to rapid overloading and potential errors, due to imperfect knowledge of server states and increased latency from coordinating load information.
Innovation Solution
Implementing a probability density function-based load balancing algorithm that biases against selecting younger servers, reducing the likelihood of focus firing by controlling the load distribution and avoiding overloading of newly added servers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If load balancers route workloads to newly added servers based on current load information, then workload distribution efficiency is improved, but newly added servers become overloaded due to focus firing
Solution Approach 1:
The system performs preliminary actions by setting an initial load threshold for newly added servers before they receive full workload routing. This preventive measure ensures that even though load balancers efficiently distribute workloads, newly added servers are protected from sudden overload by limiting their initial capacity to handle load, thus avoiding focus firing while maintaining overall distribution efficiency.
2Measurement precision
If load balancers coordinate to share perfect knowledge of server states, then routing accuracy is improved, but system latency increases
Solution Approach 1:
The system applies local quality by implementing different load threshold settings for different server groups. Newly added servers have a lower initial load threshold compared to established servers. This localized differentiation allows load balancers to make accurate routing decisions without requiring complete coordination and information sharing across all servers, thus maintaining routing precision while reducing the latency associated with system-wide coordination.
3Speed
If multiple load balancers independently route workloads, then routing speed is improved, but focus firing occurs on underloaded servers
Solution Approach 1:
The system changes the parameter of load threshold dynamically based on server age and status. Newly added servers operate with a lower initial load threshold that automatically increases over time. This parameter change allows multiple load balancers to independently and quickly route workloads without causing focus firing, as the adaptive threshold ensures newly added servers gradually accept load rather than being overwhelmed by simultaneous routing decisions.
Data Source
AI summary
Systems and methods are described for allocating requests to implement new workloads within a dynamic set of servers. Existing load balancing techniques can result in “focus firing” on new servers added to the set, since a load balancer may view a new server as underloaded. With sufficient intensity, focus firing can result in overshooting target load for the new server, and the new server in fact becoming overloaded. The present disclosure modifies selection of servers as potential targets for a workload by at least partly biasing against selection of young servers. The bias imposed can be scaled to avoid overloading new servers.