Asynchronous Network Latency Analysis via Busy Period Simulation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for determining worst case latency and backlog in asynchronous networks, such as queuing theory, network calculus, and simulation, often provide pessimistic or unreliable results, making it challenging to verify and certify performance characteristics of critical systems like flight control systems.
Innovation Solution
A system and method that utilize a processor and memory with a scheduler to determine the maximum busy period length, candidate starting times, and maximum layout for information flows, allowing for precise calculation of worst case latency and backlog by identifying the maximum busy period and updating latency and backlog values accordingly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If queuing theory is used to analyze latency and backlog, then average latency and backlog can be determined, but worst case latency and backlog cannot be determined
Solution Approach 1:
The patent segments the analysis into two distinct parts: using queuing theory to determine average latency and backlog, and using simulation methods to determine worst case latency and backlog. This segmentation allows each method to be applied where it is most effective, resolving the contradiction between measuring average values and determining worst case values.
Solution Approach 2:
The patent performs multiple rounds of simulation beyond what queuing theory alone can provide. By running numerous simulation iterations and collecting maximum latency and backlog values, the system obtains worst case scenarios that complement the average values from queuing theory, thus achieving both average and worst case measurements.
2Reliability
If network calculus is used to derive pessimistic bounds, then worst case latency bounds can be obtained, but the bounds are far from real values and resources are wasted
Solution Approach 1:
The patent changes the approach from using pessimistic analytical bounds to using simulation-based empirical measurements. By adjusting the methodology from theoretical worst case derivation to actual simulation observation, the system obtains tighter, more realistic worst case values that do not require excessive resource provisioning, thus improving resource utilization while maintaining reliability.
Solution Approach 2:
Instead of relying on pessimistic theoretical bounds, the patent creates copies of the system behavior through multiple simulation runs. These simulated copies reveal the actual worst case scenarios without the excessive conservatism of analytical bounds, allowing for more efficient resource allocation.
3Productivity
If simulation methods are used to analyze latency and backlog, then maximum values can be collected, but the results are not guaranteed to be actual worst cases
Solution Approach 1:
The patent implements a feedback mechanism where simulation results are continuously analyzed and compared. By running multiple simulation rounds and tracking the maximum observed latency and backlog, the system uses this feedback to approach the actual worst case values. The iterative nature of the simulation provides increasing confidence that the observed maxima represent true worst cases, bridging the gap between speed and reliability.
4Measurement precision
If mathematical programming is used to create high-fidelity models, then accurate latency analysis can be achieved, but it is challenging to create correct models with proper variables and parameters
Solution Approach 1:
The patent introduces simulation as an intermediary between the complex mathematical programming models and the actual system behavior. Rather than directly creating and solving complex optimization models, the system uses simulation to capture system behavior, which is then analyzed to determine latency and backlog characteristics. This intermediary approach simplifies the modeling process while maintaining accuracy.
Data Source
AI summary
A system for determining a worst case latency for a specific information flow that is part of a plurality of information flows and a worst case backlog for a specific queue that is part of a plurality of queues is disclosed. The plurality of information flows and plurality of queues are part of a configuration. The system performs operations including determining a maximum busy period length for the configuration. The operations include determining a set of candidate starting times for the configuration based on the maximum busy period length. The operations further include determining a maximum layout for a plurality of information flows within the configuration. The operations include updating the worst case latency and the worst case backlog based on the maximum layout. Finally, the operations include determining the worst case latency for the specific information flow and the worst case backlog for a specific queue.