Consolidator for Multi-Core Load Balancing Metrics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In multi-core systems managing network traffic, synchronizing metrics across multiple packet processing engines operating on independent cores is challenging due to changing metric parameters and values, leading to outdated information that affects service accuracy.
Innovation Solution
A method is implemented where a timer is established for each packet engine, and a consolidator stores and updates counter values across cores, sending consolidated metrics to ensure accurate and up-to-date information for load balancing and network traffic management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If packet engines operate independently on separate cores, then processing speed and parallelism are improved, but metric synchronization and service accuracy deteriorate
Solution Approach 1:
A consolidator component is introduced as an intermediary between packet engines on different cores. The consolidator collects metrics from each packet engine, consolidates them into a unified view, and distributes the consolidated metrics back to all packet engines. This mediator enables independent cores to share synchronized metric information without compromising their parallel processing capability.
2Reliability
If metrics are updated frequently across all packet engines, then service accuracy is improved, but system complexity and communication overhead increase
Solution Approach 1:
The consolidator merges metric collection, consolidation, and distribution functions into a single centralized component. Instead of each packet engine independently managing metric synchronization, the consolidator handles all metric updates in one place, reducing the complexity of inter-core communication and coordination logic across the system.
Solution Approach 2:
The system implements periodic metric consolidation where the consolidator collects metrics from packet engines at regular intervals and distributes updated consolidated metrics to all packet engines. This periodic action ensures metrics remain synchronized without requiring continuous communication, reducing system complexity while maintaining service accuracy.
3Measurement precision
If consolidated metrics are distributed to all packet engines, then load balancing accuracy is improved, but communication time and data exchange increase
Solution Approach 1:
The consolidator creates copies of the consolidated metric data and distributes them to all packet engines simultaneously. Each packet engine receives an identical copy of the consolidated metrics, enabling all engines to operate with the same up-to-date information without requiring continuous inter-engine communication, thus reducing communication time while maintaining load balancing accuracy.
Data Source
AI summary
Systems and methods for consolidating metrics and statistics used for load balancing by a plurality of cores of a multi-core intermediary are disclosed. A timer operating on each packet engine of each core in a multi-core system may expire. A consolidator may store, responsive to expiration of the timer, a set of counter values from each of the packet engines to a first storage location. The consolidator may send to each packet engine a message to update the set of counter values. The consolidator may, upon completion of updating the set of counter values by the packet engines, send a second message to the packet engines that includes a consolidated set of counter values determined based on the updated set of values from each packet engine. Each packet engine may establish settings and parameters for load balancing based on the consolidated set of counter values.