LogoLogo
WebsiteDIA AppRequest integration
  • DIA Oracles Overview
  • Intro to Lumina
    • How it works
    • Integrated chains
  • Data products & Tools
    • Token price feeds
  • Oracle Builder (Beta)
  • How-to guides
    • Fetch price data
      • Push based oracles
      • Pull based oracles
      • Contract Addresses
    • Run a Feeder
      • Hardware specs
      • Setting up the node
        • Docker compose
        • Docker run
        • Kubernetes
    • Stake DIA
    • Request a custom oracle
  • Reference
    • Architecture
      • Data sourcing
      • Data computation
      • Data delivery
    • Cross-chain Messaging
  • Resources
    • Audits
    • Chain information
    • Nexus Documentation
    • Community & Support
    • T&C
      • Licence Agreement
      • Contributor Covenant Code of Conduct
      • Disclaimer
Powered by GitBook
On this page
  • Cost Projections
  • Performance Metrics
  1. How-to guides
  2. Run a Feeder

Hardware specs

Hardware
Normal Usage
Peak Usage

Memory

~100 MiB

~500 MiB

CPU

~0.1 cores

~0.3 cores

Storage

Minimal

N/A

The storage is minimal because no historical data is stored on the nodes; only real-time data is pushed directly to smart contracts. These resource needs can be sufficiently met by a minimal, entry-level virtual machine from most cloud providers with the following specifications:

  • 1 vCPU

  • 1 GB RAM

  • 10 GB Storage

Cost Projections

Standalone VM

The cost of operating the decentral-feeder primarily depends on the infrastructure selected by the node operator. The following prices are estimates and may vary depending on factors such as region, additional services, network usage, and other considerations.

Cloud Provider
Monthly Costs

Hetzner (CX22)

~$4

Amazon AWS (t2.micro)

~$9

DigitalOcean (Basic Droplet)

~$5

Kubernetes Clusters

The costs of running feeder nodes in our Kubernetes clusters are generally higher compared to standalone VMs. It also depends on your setup and the additional overhead of managing Kubernetes infrastructure and other services.

Performance Metrics

  • Our nodes are configured with 20-100 websocket connections open in parallel, and a transaction is being sent every 1-20 seconds so a similar configuration would be expected

  • You can use public RPC endpoints to fetch data and submit txns onchain. However, they’re mostly good for testing purposes so we recommend setting up a private endpoint for consistent usage. Here are some of our recommendations:

PreviousRun a FeederNextSetting up the node

Last updated 1 month ago

Chainstack
Alchemy
Ankr
Quicknode