Notes on Avalanche’s Subnets

Overview

In this post, I am capturing notes from an outstanding presentation by the Ava Labs team on subnets. I highly recommend watching the video to get a better understanding of subnets. Please note I took these notes while watching the presentation; therefore they might be a bit rush and not entirely accurate.

Horizontal Versus Vertical Scaling

• Vertical vs. horizontal scaling is a big issue with the blockchain.

• The Merkle-tree used for compression can be inefficient in many ways

  • Too many writes to disk.

  • Some solutions were initially reasonable but were outdated with current use cases.

• We need to scale more than 100x and focus more on infinite scaling.

• Horizontal scaling can be a great implementation if you can isolate blockchains for specific use cases. It’s easier if ten nodes only focus on their use case instead of keeping track of the overall state.

What are subnets

• A new blockchain that runs on the avalanche ecosystem

• It takes two things to run a subnet

  • A virtual machine - Application-level logic of your blockchain. A representation of your state.

    • It contains all your transactions that manipulate your state.

  • A validator set - Provides the set of requirements needed for the blockchain.

    • Including things like KYC, hardware requirements, etc.

• On AVAX, the P-chain, C-chain, and X-chain can communicate with each other. In the future, subnets will have the capability of keeping track of only their state but also communicating with each other.

• Subnets are the “main catch” of Avax.

• Subnets allow you to deploy your own blockchain on Avalanche.

  • You need to provide the virtual machine (application rules) and the validator set (who can participate).

  • Allows you to have private and public blockchains.

  • Different virtual machines can be run on Avalanche.

    • EVM - Ethereum

    • Bitcoin’s VM

    • Solana’s VM

    • Cardano’s VM

Power of Subnets

Compounding networking effect (Reed’s Law) - The deeper the network(s) are, the more value it has.

• The power of the AVAX subnet is that it can recreate many popular blockchains (and new blockchains) and allow for interpolation. You can have transactions between subnets.

• Hypothetically speaking, you can run a bitcoin subnet on AVAX, which uses PoS instead of Nakamoto Protocol. This would allow you to have significantly faster transactions while still utilizing bitcoin.

• Avax Consensus - Avalanche will determine the next block for you utilizing their consensus model. The creator of new blockchains only has to specify the actions that need to occur after a new block is mined.

• You don’t have to worry about the language; you simply need to implement the GRPC client into your logic.

• You can have instant access to liquidity if your subnet is on Avalanche.

  • Move assets from any subnets.

• No need to build bridges. Avax has put a lot of effort into their bridge, especially since it’s a critical part of the subnet architecture.

• You can build Layer 1’s on these subnets.

• With Avax subnet, you can focus more on your application/business-specific logic.

• Existing L1s might want to deploy their virtual machines on AVAX.

• You can also have private blockchains but have them managed by the chain.

  • Regulatory compliance for banks can utilize this logic.

• Anyone can launch a subnet (permissionless); you don’t need to lock up funds. The only requirement is to serve as a validator of the P-chain.

• You might want to deploy a subnet for individual Dapp, similar to a microservice model.

• Deployment is super easy cause it’s all done through a JSON file.

  • In the future, there might be a frontend that allows users to interact with a frontend to create a subnet.

• If you run a subnet, you can choose what token is used for transactions and gas.

  • You can pay validators in the chosen token.

How to Leverage Subnets

• Knowing Golang is a benefit when working with AVAX.

• Messing around with the EVM JSON config is a great way to experiment and start.

• Try to build a real blockchain on the subnet.

• Try to build a non-EVM blockchain on the subnet.

• The hardest thing might be to create a new VM from scratch.

  • Although the EVM is great, if history has taught us anything, everything gets replaced. Try to create the successor of the EVM.