ANALYSIS OF THE PERFORMANCE OF A MONITORING SYSTEM BASED ON GNMI STREAMING TELEMETRY WITH CONTAINERLAB NETWORK SIMULATION
The rapid growth of the internet has an impact on the development of digital services. This surge in demand has created opportunities for digital services industry players. Despite its positive impact, the growth of the internet has also created technical challenges. In the context of dealing wit...
Saved in:
Main Author: | |
---|---|
Format: | Theses |
Language: | Indonesia |
Online Access: | https://digilib.itb.ac.id/gdl/view/79085 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Institution: | Institut Teknologi Bandung |
Language: | Indonesia |
Summary: | The rapid growth of the internet has an impact on the development of digital
services. This surge in demand has created opportunities for digital services
industry players. Despite its positive impact, the growth of the internet has also
created technical challenges. In the context of dealing with growing data traffic,
resource monitoring is a crucial factor. One of the newest ways to monitor this
resource is to use the gNMI Streaming Telemetry system. Even though it seems
superior to existing protocols, further studies in implementing the Streaming
Telemetry system need to be studied. This paper is focused on investigating the
trade-offs and performance of gNMI Streaming Telemetry. Design and simulation
were carried out using Docker-based networking lab tools, namely Containerlab.
Integration of the monitoring system and network topology is also carried out in
the Docker-based simulation. Observation results from each protocol show that the
activity of taking metrics by the monitoring system does not have a significant effect
on network conditions. This is indicated by the low average network latency and
almost uniform throughput except in lossy and congestion conditions. Observations
in the simulation also show that the gNMI monitoring system uses I/O resources
more intensively than other protocols. The research also discusses the integration
of gNMI Streaming Telemetry and log monitoring, which shows an increase in
memory usage of 70MB and Disk I/O resources increasing by 33%. In this research,
indications were also found of increasing CPU utilization by the gNMI monitoring
system by 50% of the mode data recorded in the observations. |
---|