Skip to content

Roadmap Backends

Jump to bottom
Francois-Rene Rideau edited this page Oct 13, 2021 · 2 revisions

Support multiple blockchains.

See costs at the end.

EVM

Plutus

  • (2020) Work on Plutus Application Framework mock blockchain
  • (2021Q3) Run on production Cardano blockchain
  • (2021Q4) Compile directly to Plutus Core

Nervos

Bitcoin Script

Each Bitcoin fork has its own variant of Bitcoin Script, and while none of them is an agreeable virtual machine to target, it is possible to write a compiler backend that targets most of them. However, there are many issues to overcome

  • The variant of Bitcoin Script in BTC was wantonly enucleated (this was one of the reasons of the schism with BCH). Notably, a transaction can't verify any commitment to the outputs, which severely limits how interactions can do dynamic price computations. It is possible to e.g. split the contract in one sub-contract per bit of the price, each with its redundant UTXO, but this would induce a large multiplier (10x-50x) to scaling issues and transaction fees that are already highly problematic.
  • BCH, the Liquid Network (Blockstream's side-chain for BTC), and many other forks or clones use variants of Bitcoin that can verify output commitments, and thus be used in semi-sensible contracts. But even then, just the "administrative" overhead to maintaining contract state consumes most of the resource limits of each transaction, which means that a lot of optimizations have to go into the compiler just to fit useful chunks of computations into a transaction. This is a lot of work, and the transaction fees will still be high, so even then most users will want to minimize their use of such solutions. Thus, an expensive task for a reduced market, clearly not a low-hanging fruit.
  • Hopefully, at some point, some leaders in the Bitcoin community will fund such work and/or amend their chains to be more usable, but these chains tend to have governance issues, and it may take years for progress to be visible.
  • (2022?) Bitcoin Cash A112
  • Bitcoin
  • Liquid Network
  • Decred
  • ...

RiscV

  • Nervos

Solana

  • Solana

WebAsm

  • PolkaDot
  • Fluence

Michelson

  • Tezos

Pact

  • Kadena

Scilla

  • Zilliqa

Algorand

  • Algorand

More?

Centralized solutions

Many digital assets, being centrally issued and backed, may as well be centrally managed for much reduced fees. We shall support direct execution of Glow DApps against such platforms, easy and secure deployment of these platforms for central issuers of digital assets.

Costs

How difficult is it to support another blockchain? It depends.

  • A generic EVM blockchain? For $25K/year, MuKn will not just add a few lines in a config file, but add a test job on our CI toolchain, and making sure the network keeps working. You also pay for the marketing presence.
  • Same as above plus but with additional modifications to the EVM? Depending on the modifications, probably one month of work and $25K to $50K to initially support the chain, plus between $30K and $100K a year to make sure things keep working as your and our code evolve.
  • A non-EVM blockchain designed for Smart Contracts, such as EOS, Tezos, Nervos, Algorand, etc.? Two months and $50K for an initial prototype, then depending on issues found or not, $0 to $300K to make it an actual product. Then $150K to $300K a year to keep it running as your and our code evolve.
  • A Bitcoin clone or fork? Assuming it has the opcodes missing in BTC, that allow to write a smart contract (like BCH's OP_CHECKDATASIG or blockstream's OP_CHECKSIGFROMSTACK, or Bitcoin Unlimited's OP_DATASIGVERIFY, maybe OP_CHECKTEMPLATEVERIFY), then it's 6 months and $300K to work around all the quirks and limitations of the script, and then $300 to $1M a year to maintain the code and keep making the performance improvements.
  • Current BTC network without the required opcode updates? It's actually still possible, but will cost $1M over a year just for the initial development, to use the advanced tricks required to make it work at all, at which point fees in the general case may be 10x-50x higher than the already high fees above. Then $500K to $3M per year to keep it running.

Glow Wiki

Home

Roadmap

Design

Bibliography

Internals

Other

Clone this wiki locally