Vac R&D Roadmaps and Reports

Search

SearchSearch

Codex Scaling

Oct 08, 2024, 3 min read

vac:dst:codex:codex-scaling

Description

Use real world testing, theoretical analysis and simulation to determine and improve Codex’s scaling properties.

Find the limits of Codex’s capabilities and measure them in different scenarios.

We will allow Codex to scale to support large scale use cases, test how it behaves in large 100TB+ testnet deployments and in various deployment setups, and we will help make Codex more scalable in the first place.

We will support the Conduit of Expertise narrative directly by providing valuable insights to Codex and the ability to theorise, reason about, test, measure and improve the performance, stability and scalability of Codex.

These efforts will contribute in these ways to the Conduit of Expertise narrative:

  • Accelerate adoption and development and productising of Codex by providing support to the Codex team in the form of real world testing to improve their efficiency and effectiveness in building a better product.

  • Improve the RFC culture by allowing for faster and easier development of RFCs with the aid of rapidly accelerated insights into how an RFC in development will perform as it’s being expanded and going through the draft process.

  • Allow easier post-mortem analysis of the success or relative performance of a given RFC - does this change use more or less bandwidth? Did it improve things? Seeing the effects of changes at scale allows for a greater ability to usefully wrap up work on and conclude an RFC process and document and absorb what we learned in the process into further improvements.

Further, we will contribute both directly and indirectly to the Premier Research destination narrative by helping Codex build a stable base on which other research and interesting use cases can be built.

Task List

Deploy Base Capacity

Description

Deploy a large set of base capacity to the Codex testnet and keep it online, stable and prevented from losing data where possible.

It will consist of 50x nodes with 10xTB of data each.

Deliverables

  • Helm chart adapted to Vaclab and used to deploy the nodes.
  • 50x nodes running and adopted into the testnet.
  • Downloads/uploads tested and working for at least 3 selected nodes.
  • Ongoing monitoring (not a one time thing)
  • 500TB of overall capacity provided to the network

How Fast Is Codex?

Description

Related to Codex Comparison, we simply want to find out fast Codex is, at various things under different kinds of stress and load.

We will use the Base Capacity.

We will test and compare the following:

  • Upload speed (1 client)
  • Download speed
  • Time to first byte
  • Time to 50%
  • Time to 90%
  • Time to 100

We would also like to collect all data from the items in this matrix:

Benchmark conditions:

  • total size: 2, 8, 16, 32
  • seeders: 1, 2, 4, 8, 16
  • file size: 100 MB, 1 GB, 5 GB

Deliverables

  • Reports from how each item in the matrix performed.
  • A general writeup

Graph View

Backlinks