Skip to content

Latest commit

 

History

History
181 lines (132 loc) · 9.44 KB

cost-comparing-research-rust-cdt.md

File metadata and controls

181 lines (132 loc) · 9.44 KB
Raw

EOS Network Foundation Grant Proposal

  • Project Name: Research for building Rust Antelope CDT
  • Team Name: Lowkey codes
  • EOS Payment Address: delightlabs1
  • Level: 1
  • Pomelo Grant(s): N/A
  • Project is Open-Source: Yes
  • Project was part of Token sale: No
  • Repository where Project resides: N/A

Contact

Project Overview

Overview

  • Name: Research for building Rust Antelope CDT
  • Brief Description: Checking the most cost-effective plan among the cases we thought of is mandatory for building Rust Antelope CDT.
  • Relationship to EOSIO: Many developers using other languages besides C++ will be able to join the Antelope ecosystem and become its builders. We set the first target to Rust.
  • Reason for Interest: Lack of language support despite the extremely fast processing speed leaves something to be desired, and we anticipate that modern language support & language diversity help improve this point and expand builder pool. Additionally, in terms of our interests, we would like to contribute to technology-focused areas.

Project Details

The project will investigate the following two parts in order to verify the suitable direction to implement:

  1. Sufficient compatibility between EOS VM and wasm binary from Wasmer-based compiler
    • Rust antelope.cdt API implementation
  2. Lack of compatibility between EOS VM and wasm binary from Wasmer-based compiler exists
    • Developing compiler first -> Rust antelope.cdt API implementation
    • Or Wasmer VM integration on Antelope

This research has been written as a separated proposal since the follow-up plan will be quite different by the result.

Ecosystem Fit

This implementation allows Antelope to extend one more contract language, Rust, and this would be the cornerstone of the multiple language support in the Antelope ecosystem. As WASM can be built with Golang and Javascript recently, we look forward Antelope to embracing more other languages' users.

Team

Team members

  • Team Leader: Yeon HWANG
  • Joon LEE
  • Young LEE
  • Derick MOON
  • Sooyoung HA
  • Bryan RHEE
  • Joowon YUN
  • Maro KIM

Legal Structure

  • Registered Legal Entity: Lowkey codes
  • Registered Address: Hyecheon Bldg 1126-7 (11th flr), 354 Gangnam-daero, Gangnam, Seoul, SOUTH KOREA

Team Experience

  • Before blockchain, well experience in OS & emulator development (Tizen of Samsung)
  • Much experience in validating Cosmos SDK-based projects & ETH layer 2 blockchains
  • Experience in blockchain core development (connect between Tendermint & WASM execution engine)
  • 1st dApp on Terra(now Terra classic): Terraswap - enough experience of WASM-based smart contract development & action processing
  • Mainnet technical partner of Xpla (former C2X)
  • EOSIO-based project-building experience (Polaris of Chain Partners, mentor & participant of EOSIO hackathon)

Team Org Repos

https://github.com/DELIGHT-LABS

Team Member Repos

Team LinkedIn Profiles (if available)

Development Status

Description of the expected scenarios

┌──────────────┐             ┌──────────────┐            ┌─────────────────┐
│              │  eosio-llvm │              │            │      API        │
│   EOSIO C++  │  ───────►   │      IR      │   ───────► │    EOSVM IR     │
│              │             │              │            │ Interface set   │
└──────────────┘             └──────────────┘            └─────────────────┘

┌──────────────┐  Wasmer's   ┌──────────────┐            ┌─────────────────┐
│              │ rust-llvm   │              │            │      API'       │
│ Wasmer Rust  │  ───────►   │     IR'      │   ───────► │   Wasmer IR     │
│              │             │              │            │ Interface set   │
└──────────────┘             └──────────────┘            └─────────────────┘

NOTE: There is no official general Rust LLVM. This standardization project is still in incubation status.

Assumptions: For cost-effective building, we want to build from the many works of Wasmer, although Antelope doesn't need to mount Wamser.

Cases

  1. If IR ≅ IR' and API ≅ API'
    • We can directly start from Wasmer's implementation
      • Implement Antelope-specific data structure and API of eosio.cdt (Like table, vector, etc.)
  2. Else if IR ≅ IR' and API != API'
    • Implement a compiler, which works for adjusting WASM API interface from Wasmer spec to EOSVM spec
    • After then, start (1)
  3. Else (IR != IR' and API != API')
    • Mount Wasmer on Antelope would be the cheapest
    • Mount Wasmer, make compatible with the existing state DB, and work on (1)
    • Draft proposal of this case

Why Wasmer?

Performance of Wasmer is placed in the top tier. See benchmark 2021. Also, it has the biggest ecosystem among WASM VM and supports many languages accordingly.

Development Roadmap

Milestone Summary

  • Total Estimated Duration: 1 month
  • Full-Time Equivalent (FTE): 2 FTE
  • Total Costs: 10,000 USD

Milestone 1 — Research the difference between EOSVM and Wasmer

  • Estimated duration: 1 month
  • FTE: 2
  • Costs: 10,000 USD
ID Deliverable Specification
0a. License MIT
1 Research report 1. EOSVM API analysis
Analyze the interacting APIs with EOSVM and estimate the job size of developing the interaction layer

2. Antelope code structure research
Check does the code need to be changed & where would need to be changed

3. Feasibility check
Implement a small PoC layer including with a couple of API interaction & triggering contract.

4. (OPTIONAL) Simulation with DEX contract
Try to simulate porting the Cosmwasm DEX contract on EOSVM

Future Plans

There will be a follow-up proposal for supporting Rust CDT according to this research result.

Additional Information

How did you hear about the Grants Program? Personal recommendation

Related experiences