CosmWasm Smart Contracts to Solana Programs

Learn the difference between CosmWasm and Solana smart contracts.

CosmWasm

• Execution Environment: Runs on the Cosmos SDK, designed for interoperability between blockchains.
• State Management: Uses cosmwasm_storage for storing and retrieving state.
• Entry Points: Separate entry points for instantiate, execute, and query.
• Messages: Uses InstantiateMsg, ExecuteMsg, and QueryMsg to define contract interactions.

Solana

• Execution Environment: Runs on the Solana blockchain, optimized for high throughput and low latency.
• State Management: Primarily uses Borsh for serialization and deserialization of state stored in accounts. Solana recommends Borsh for its efficiency, but developers can use other serialization formats like bincode if desired.
• Entry Point: A single entry point (process_instruction) handles all instructions.
• Instructions: Uses custom-defined instructions encoded typically as byte arrays, which are parsed within process_instruction to determine the specific operations to perform. Instructions must be explicitly defined and handled within the smart contract.

State Management #

In CosmWasm, state is stored in the smart contract's storage, which is managed by the Cosmos SDK. This contract interacts with this storage via helper functions provided by cosmwasm-storage. This storage is part of the blockchain state and is specific to each contract instance. State variables are typically defined in a state.rs file and are serialized/deserialized using serde. State management functions are used within the contract logic to read from and write to the storage.

In contrast, Solana programs manage state by interacting with the data of specific accounts that are passed into the program during execution. These accounts can be created and assigned by the System Program and are allocated a certain amount of space for storing data. The state within these accounts is serialized and deserialized using libraries like Borsh. Solana programs are responsible for interpreting the account data as state information, writing to and reading from these accounts directly in the program logic.

The core concept of accounts on Solana is crucial to understanding state management and effectively designing Solana programs. Each account on Solana can be associated with a specific program (smart contract), and only the owning program can modify its state. This model differs significantly from many other blockchain platforms where the state is managed more directly through the smart contract itself. For detailed information on the Solana Account Model, refer to the Solana documentation here.

Key Differences #

1. Storage Abstraction:
   • CosmWasm: Uses abstractions like cosmwasm_storage to manage state. The contract interacts with a key-value storage system provided by the Cosmos SDK.
   • Solana: Directly reads from and writes to account data, which is part of the account's state on the Solana blockchain.
2. Serialization:
   • CosmWasm: Uses a mix of JSON for interoperability and human readability, and binary formats like bincode for state storage.
   • Solana: Typically uses Borsh for all serialization purposes to ensure speed and consistency in a high throughput environment, but have the option to roll your own if you choose to.
3. State Location:
   • CosmWasm: State is stored in the contract's storage, managed by the Cosmos SDK.
   • Solana: State is stored in account data, which is passed to the program during execution.
4. Access Patterns:
   • CosmWasm: State is accessed via storage APIs provided by CosmWasm.
   • Solana: State is accessed by directly manipulating the data of accounts.

Converting CosmWasm State Management to Solana #

Step 1: Define Account State #

In Solana, state is stored in accounts. Define the state structure and manage serialization using Borsh.

CosmWasm State Definition

#[derive(Serialize, Deserialize, Clone, Debug, PartialEq)]
pub struct State {
    pub count: u32,
}

Solana State Definition

use borsh::{BorshDeserialize, BorshSerialize};
 
#[derive(BorshSerialize, BorshDeserialize, Debug)]
pub struct CounterState {
    pub count: u32,
}

Step 2: Reading and Writing State #

In Solana, State is read from and written to account data in the program logic.

use solana_program::{
    account_info::AccountInfo, entrypoint::ProgramResult, program_error::ProgramError, pubkey::Pubkey,
};
use borsh::BorshDeserialize;
use std::convert::TryInto;
 
pub fn process_increment(accounts: &[AccountInfo], program_id: &Pubkey) -> ProgramResult {
    let account_info_iter = &mut accounts.iter();
    let counter_account = next_account_info(account_info_iter)?;
 
    if counter_account.owner != program_id {
        return Err(ProgramError::IncorrectProgramId);
    }
 
    let mut counter_data = Counter::try_from_slice(&counter_account.data.borrow())?;
    counter_data.count += 1;
    counter_data.serialize(&mut &mut counter_account.data.borrow_mut()[..])?;
 
    Ok(())
}

Entry Points #

CosmWasm provides a modular approach with separate entry points (instantiate, execute, query) for different types of operations. Solana uses a single entry point (process_instruction) for all instruction types, which offers fine-grained control.

Key Differences #

1. Entry point structure
   • In CosmWasm, each entry point is defined as a separate function, often using the #[entry_point] attribute in Rust.
   • In Solana, the program internally dispatches the instruction to the appropriate handler based on teh instruction data.
2. Serialization and Deserialization
   • CosmWasm contracts use JSON for message serialization and deserialization, which makes it easier to understand and debug but can introduce overhead
   • Solana Programs use Binary Serialization for instructions, which is more efficient in terms of performance and storage.
3. Error Handling
   • CosmWasm typically uses standard Rust error handling methods and the errors are returned as part of the result
   • Solana has a specific set of program errors and error handling is closely tied to instruction processing. Errors must be propagated correctly to ensure proper transaction behavior.

Converting CosmWasm Entry Points to Solana #

CosmWasm EntryPoint Example

use cosmwasm_std::{entry_point, DepsMut, Env, MessageInfo, Response, StdResult};
 
#[entry_point]
pub fn instantiate(
    deps: DepsMut,
    env: Env,
    info: MessageInfo,
    msg: InstantiateMsg,
) -> StdResult<Response> {
    // Initialization logic
}
 
#[entry_point]
pub fn execute(
    deps: DepsMut,
    env: Env,
    info: MessageInfo,
    msg: ExecuteMsg,
) -> StdResult<Response> {
    match msg {
        ExecuteMsg::Increment {} => execute_increment(deps, env, info),
        ExecuteMsg::Reset { count } => execute_reset(deps, env, info, count),
    }
}
 
#[entry_point]
pub fn query(deps: Deps, env: Env, msg: QueryMsg) -> StdResult<Binary> {
    match msg {
        QueryMsg::GetCount {} => query_count(deps),
    }
}

Solana Entry Point Example

use solana_program::{
    account_info::AccountInfo,
    entrypoint,
    entrypoint::ProgramResult,
    pubkey::Pubkey,
    program_error::ProgramError,
};
use borsh::BorshDeserialize;
 
entrypoint!(process_instruction);
 
fn process_instruction(
    program_id: &Pubkey,
    accounts: &[AccountInfo],
    instruction_data: &[u8],
) -> ProgramResult {
    let instruction = CounterInstruction::try_from_slice(instruction_data)
        .map_err(|_| ProgramError::InvalidInstructionData)?;
 
    match instruction {
        CounterInstruction::Initialize { count } => {
            process_initialize(accounts, count, program_id)
        }
        CounterInstruction::Increment => process_increment(accounts, program_id),
        CounterInstruction::Reset { count } => process_reset(accounts, count, program_id),
    }
}

Messages vs. Instruction Handling #

Understanding the key differences between CosmWasm messages and Solana program instruction handling is crucial for developers transitioning between these two ecosystems

Key Differences #

CosmWasm smart contracts use a message-based architecture to handle different types of interactions. These messages define the inputs and operations that a contract can perform.

1. Message Types:

   • InstantiateMsg: Defines parameters for contract instantiation.
   • ExecuteMsg: Defines parameters for executing various contract functions.
   • QueryMsg: Defines parameters for querying the contract state without changing it.

2. Handling Messages:

   • Separate entry points for different message types: instantiate, execute, and query.
   • Each entry point processes its specific message type and performs the corresponding logic.

Solana programs use instructions to define operations that an on-chain program can perform. These instructions are more granular and low-level compared to CosmWasm messages.

1. Instruction Definition:

   • Instructions are defined using custom data structures.
   • Typically serialized and deserialized using Borsh.

2. Account Handling:

   • Instructions include references to accounts that the program will read from or write to.
   • Ensures the program has the necessary access to operate on the blockchain state.

Converting Messages to Instructions #

Step 1: Define Messages as Instructions #

In CosmWasm, messages are defined using structs and enums. In Solana, similar structs and enums will be used to define instructions.

CosmWasm Message Definition:

use serde::{Deserialize, Serialize};
 
#[derive(Serialize, Deserialize, Clone, Debug, PartialEq)]
pub struct InstantiateMsg {
    pub count: u32,
}
 
#[derive(Serialize, Deserialize, Clone, Debug, PartialEq)]
pub enum ExecuteMsg {
    Increment {},
    Reset { count: u32 },
}
 
#[derive(Serialize, Deserialize, Clone, Debug, PartialEq)]
pub struct QueryMsg {
    pub get_count: {},
}

Solana Instruction Definitions:

use borsh::{BorshDeserialize, BorshSerialize};
 
#[derive(BorshSerialize, BorshDeserialize, Debug)]
pub enum CounterInstruction {
    Initialize { count: u32 },
    Increment,
    Reset { count: u32 },
}

Step 2: Implement Instruction Handling #

Create handlers for each instruction in Solana. This involves writing the logic to process each instruction and update the account state accordingly. This is very similar to handling Execute and Query messages in a CosmWasm smart contract.

Solana Instruction Handling:

use solana_program::{
    account_info::AccountInfo, entrypoint::ProgramResult, program_error::ProgramError,
    pubkey::Pubkey,
};
use borsh::{BorshDeserialize, BorshSerialize};
use crate::state::CounterState;
 
pub fn process_initialize(
    accounts: &[AccountInfo],
    count: u32,
    program_id: &Pubkey,
) -> ProgramResult {
    let account_info_iter = &mut accounts.iter();
    let counter_account = next_account_info(account_info_iter)?;
 
    // Ensure the account is owned by the program
    if counter_account.owner != program_id {
        return Err(ProgramError::IncorrectProgramId);
    }
 
    let mut counter_data = CounterState { count };
 
    // Serialize the state into the account
    counter_data.serialize(&mut &mut counter_account.data.borrow_mut()[..])?;
 
    Ok(())
}
 
pub fn process_increment(accounts: &[AccountInfo], program_id: &Pubkey) -> ProgramResult {
    let account_info_iter = &mut accounts.iter();
    let counter_account = next_account_info(account_info_iter)?;
 
    if counter_account.owner != program_id {
        return Err(ProgramError::IncorrectProgramId);
    }
 
    // Deserialize the state
    let mut counter_data = CounterState::try_from_slice(&counter_account.data.borrow())?;
 
    // Update the count
    counter_data.count += 1;
 
    // Serialize the updated state back into the account
    counter_data.serialize(&mut &mut counter_account.data.borrow_mut()[..])?;
 
    Ok(())
}
 
pub fn process_reset(
    accounts: &[AccountInfo],
    count: u32,
    program_id: &Pubkey,
) -> ProgramResult {
    let account_info_iter = &mut accounts.iter();
    let counter_account = next_account_info(account_info_iter)?;
 
    if counter_account.owner != program_id {
        return Err(ProgramError::IncorrectProgramId);
    }
 
    let mut counter_data = CounterState::try_from_slice(&counter_account.data.borrow())?;
 
    counter_data.count = count;
 
    counter_data.serialize(&mut &mut counter_account.data.borrow_mut()[..])?;
 
    Ok(())
}

Solana Program Advantages #

1. Performance Efficiency:
   • Solana’s binary instruction data and direct account manipulation provide high performance and low latency.
   • This is critical for high-throughput applications like decentralized exchanges (DEXes) and other performance-sensitive use cases.
2. Fine-Grained Control:
   • The instruction-based approach offers fine-grained control over program execution and state management.
   • Developers can optimize their programs at a low level, potentially leading to more efficient implementations.
3. Flexibility:
   • Custom instructions can be designed to fit any specific use case, providing flexibility for advanced and complex operations.
   • This makes it suitable for developers who need to implement highly specialized logic.

In conclusion, Solana is ideal for applications that require high performance, low latency, and fine-grained control over execution. It’s better suited for developers comfortable with lower-level programming and those who need to optimize for specific use cases.