Time-Triggered Network Communication Bandwidth Utilization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current network standards, such as IEEE 802.1Q-2011, limit the number of real-time messages that can be sent in a computer network without increasing latency, restricting bandwidth utilization and the number of messages that can be transmitted effectively.
Innovation Solution
Implementing time-triggered communication where messages are sent periodically according to a defined schedule, with phase-shifted transmission times and integral multiple periods of CM time intervals, ensuring non-overlapping intervals and shared bandwidth with real-time messages within the limited real-time bandwidth.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the number of real-time messages is increased in the network, then bandwidth utilization is improved, but latency increases
Solution Approach 1:
The patent implements time-triggered communication where messages are transmitted in periodic Class Measurement (CM) time intervals. Each computing node sends messages at predetermined periodic intervals rather than continuously or randomly, which allows the system to achieve high bandwidth utilization while maintaining bounded and predictable latency. The periodic transmission pattern ensures that network resources are efficiently utilized without causing unbounded delay accumulation.
Solution Approach 2:
The patent employs predetermined schedules that are established in advance for each computing node, specifying exactly when messages should be transmitted within each CM time interval. This preliminary scheduling allows the system to pre-allocate bandwidth and avoid contention, ensuring that latency remains predictable and bounded while maximizing the number of messages that can be transmitted within the available bandwidth.
2Productivity
If more computing nodes transmit messages simultaneously, then network activity increases, but message collisions and latency increase
Solution Approach 1:
The patent segments the transmission time of each computing node within a CM time interval by assigning predetermined, non-overlapping time slots to each node through scheduling. This temporal segmentation allows multiple computing nodes to transmit messages simultaneously in different time slots without causing collisions, thereby increasing overall network activity while keeping transmission delays bounded and predictable.
3Loss of time
If bandwidth for real-time messages is limited to a defined real-time bandwidth, then latency is controlled, but the number of transmittable messages is restricted
Solution Approach 1:
The patent uses periodic CM time intervals to efficiently utilize the limited real-time bandwidth. By organizing message transmissions into periodic intervals with predetermined schedules, the system maximizes the number of messages that can be transmitted within the bandwidth constraints, achieving both latency control and high message throughput.
Solution Approach 2:
The patent implements dynamic scheduling that adapts to the specific bandwidth limitations and message requirements of each computing node. The predetermined schedules are designed to dynamically allocate the limited real-time bandwidth across multiple nodes and message types, ensuring optimal utilization while maintaining bounded latency for all real-time communications.
Data Source
Figure 1~2
Figure 3~5
Figure 6~8
AI summary
The invention relates to a method for transmitting communications in a computer network consisting of compute nodes that are interconnected via active components. The compute nodes exchange real-time communications (402), wherein the real-time communications (402) are assigned defined CM-time intervals (CMI) of constant duration. The bandwidth available for real-time communications (402) within a CM time interval (CMI) is limited to a defined real-time bandwidth. Further, time-controlled communications (401, 403) are sent from compute nodes, wherein the time-controlled communications (401, 403) are sent periodically from the compute nodes. The cycle duration of the time-controlled communications (401, 403) is a function of the duration of the CM-time intervals and the transmit times of time-controlled communications (401, 403) that are sent from different compute nodes to the same active components are out-of-phase with respect to each other, such that time-controlled communications (401, 403) from different compute nodes are received in the active components at different times (601, 602). In addition, the sum of the bandwidths occupied by time-controlled communications (401, 403) and by real-time communications (402) within a CM-time interval (CMI) does not exceed the value for the real-time bandwidth.