Session Load Balancer Traffic Policies for Heterogeneous Session Controllers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current load balancing techniques, such as round robin algorithms, fail to account for heterogeneous processing capabilities among next-hop entities, leading to overloading of slower nodes and incorrect traffic distribution due to stale CPU utilization data, resulting in rejected requests and dropped packets.
Innovation Solution
Implementing intelligent load balancer traffic policies that consider specific attributes of each session controller, such as host throttle rate and maximum signaling rate, to dynamically distribute traffic and prevent overloading, while excluding nodes with high CPU utilization from receiving new sessions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If round robin algorithm is used for traffic distribution, then implementation simplicity is improved, but traffic distribution fairness deteriorates due to heterogeneous processing capabilities of next-hop entities
Solution Approach 1:
The patent applies local quality by transitioning from uniform round robin distribution to heterogeneous load balancing that assigns different weights to different next-hop entities based on their specific processing capabilities. Each entity receives traffic according to its local characteristics (CPU capacity, current load), creating a non-uniform distribution pattern that matches actual capabilities.
Solution Approach 2:
The patent changes the distribution parameter from fixed equal weights to dynamic weights based on CPU utilization metrics. The load balancer continuously monitors CPU usage and adjusts traffic distribution parameters accordingly, allowing faster entities to handle more traffic while slower ones receive less, thereby achieving fairness through parameter adaptation.
2Use of energy by moving object
If CPU utilization metrics are collected at predetermined time intervals, then system overhead is reduced, but data freshness deteriorates leading to stale data decisions
Solution Approach 1:
The patent applies dynamics by making the metric collection interval adaptive rather than fixed. The load balancer dynamically adjusts the timing of CPU utilization queries based on current system conditions and traffic patterns, collecting metrics more frequently when load changes are detected and less frequently during stable periods, thus balancing freshness with overhead.
Solution Approach 2:
The patent implements feedback mechanisms where the load balancer monitors CPU utilization metrics and uses this information to adjust traffic distribution in real-time. The system continuously receives feedback about entity performance and dynamically modifies distribution decisions, ensuring that even with periodic sampling, the most current available data drives load balancing choices.
3Productivity
If traffic is distributed based on lowest CPU utilization rate, then load balancing efficiency is improved, but request rejection increases when entities are actually busy due to stale data
Solution Approach 1:
The patent applies preliminary action by having the load balancer proactively query CPU utilization metrics before making traffic distribution decisions. Rather than relying on potentially stale periodic updates alone, the system performs preliminary checks to ensure the selected next-hop entity is actually capable of handling new traffic, preventing request rejections due to outdated information.
Solution Approach 2:
The patent uses feedback loops where the load balancer continuously monitors CPU utilization and adjusts traffic distribution based on real-time conditions. When an entity's CPU usage exceeds thresholds or changes significantly, the system receives feedback and redistributes traffic accordingly, ensuring that efficiency gains from load balancing do not come at the cost of rejecting requests to overloaded entities.
Data Source
AI summary
The subject matter herein includes methods, systems, and computer readable media for implementing load balancer traffic policies. An exemplary method may be performed at a session load balancer (SLB) and may include receiving, at the SLB, one or more policy attributes associated with each session controller (SC) in a cluster of SCs. The method may further include configuring, at the SLB, a load balancer traffic policy based on the one or more policy attributes received from each SC in the cluster of SCs, where the load balancer traffic policy includes a distribution rate associated with each SC in the cluster of SCs. The method may further include receiving, at the SLB, a registration request for establishing a session with a SC, determining a destination SC using the load balancer traffic policy, and forwarding the registration request to the destination SC.