IOAM Message Sampling for Network Bandwidth and Server Load
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing IOAM technology in data center networks faces issues of excessive network bandwidth occupation due to IOAM information being added to all messages, leading to reduced effective bandwidth and payload utilization, as well as server processing capability limitations during high-rate message output.
Innovation Solution
A message sampling method in the IOAM mechanism where an encapsulation node samples and encapsulates only target messages based on a pre-configured mode, reducing network load and server processing requirements by executing the IOAM function only on selected messages.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If IOAM function is executed on all messages, then network visualization capability is improved, but network bandwidth occupation increases and effective bandwidth decreases
Solution Approach 1:
The patent applies partial action by executing the IOAM function only on sampled messages rather than all messages. The encapsulation node samples messages at a predetermined rate and applies IOAM encapsulation only to the sampled subset, thereby maintaining network visualization capability while significantly reducing bandwidth occupation and overhead.
Solution Approach 2:
The patent segments the message stream into sampled and non-sampled portions. The encapsulation node identifies and processes only the sampled messages with IOAM encapsulation, while leaving non-sampled messages to traverse the network without additional overhead, thus dividing the processing load and bandwidth consumption.
2Loss of information
If IOAM information is added to all messages, then network monitoring completeness is improved, but payload utilization rate decreases
Solution Approach 1:
The patent applies partial action by adding IOAM information only to sampled messages rather than all messages. This maintains sufficient monitoring capability for network analysis while preserving payload utilization rate by avoiding unnecessary overhead on the majority of non-sampled messages.
3Reliability
If IOAM function is executed on all messages, then network operation data acquisition is improved, but server processing capability requirements increase
Solution Approach 1:
The patent applies partial action by processing only sampled messages through the IOAM function. The encapsulation node samples messages at a predetermined rate and applies IOAM encapsulation only to the sampled subset, significantly reducing the processing load on servers while maintaining network operation data acquisition capability.
Solution Approach 2:
The patent segments the processing workload by separating sampled messages (requiring IOAM processing) from non-sampled messages (bypassing IOAM processing). This division reduces the overall processing complexity and server capability requirements while maintaining adequate monitoring coverage.
4Reliability
If IOAM function is executed on all messages, then network fault detection capability is improved, but network load increases
Solution Approach 1:
The patent applies partial action by executing the IOAM function only on sampled messages rather than all messages. This maintains network fault detection capability through statistical sampling while significantly reducing network load and overhead.
Data Source
Figure 1~2
Figure 3~5
Figure 6~7
AI summary
Disclosed by the embodiments of the present application are a message sampling method in an in-situ operation administration and maintenance (IOAM) mechanism, a method for unpacking data on the basis of the message sampling method in the IOAM mechanism, a packaging node, an unpacking node, a message sampling system and a storage medium, the message sampling method comprising: receiving transmission data; and according to a configured sampling mode, carrying out an IOAM function on a target message matching a sampling mode in the transmission data.