Celestia Light Node performance analysis

Description.

In this article we performed an analysis of Light Node effectiveness depending on hardware.

Environment.

Provider: https://oneprovider.com/

Location: Dallas (used same location to exclude location factor)

VM 1: Core(s) • 30GB SSD Storage • 1GB Memory

VM 2: Core(s) • 40GB SSD Storage • 2GB Memory

VM 3: Core(s) • 60GB SSD Storage • 3GB Memory

Tools used for testing:

Ansible:

I prepared an Ansible playbook to deploy Light Node. Here is link to my Github repo https://github.com/krompets/celestia-light-node-playbook

Usage:

1. Clone project
   # git clone https://github.com/krompets/celestia-light-node-playbook.git
2. Go to project dir
   # cd celestia-light-node-playbook
3. Fill out hosts file with your hosts
4. Deploy Light Node:
   # ansible-playbook -i hosts playbook.yml

Read here https://github.com/krompets/celestia-light-node-playbook/blob/main/README.md

Grafana:

I installed Grafana and Prometheus docker containers to get hardware usage statistics

Tiascan:

For monitoring the performance I used https://tiascan.com/ as it is developed by Celestia and I can 100% trust.

Results.

• VM 1: Core(s) • 30GB SSD Storage • 1GB Memory

• VM 1: 2 Core(s) • 40GB SSD Storage • 2GB Memory

• VM 3: 3 Core(s) • 60GB SSD Storage • 3GB Memory

Conclusion:

• As we see from results we have bad uptime score on hardware VM-1 (Core(s) • 30GB SSD Storage • 1GB Memory).
• On VM-2 and VM-3 the result is perfect. So we have to use 2+ Core(s) • 40GB+ SSD Storage • 2GB+ Memory to have good performance.