HTTP batch request VS multicall contract

This article is about two approaches many developers believe can help them save on RPC consumptions: batch request and multicall smart contract.

In this article, we will explore how to use them and compare their performances.

HTTP batch request

HTTP batch request is a feature most Ethereum clients support, for example, Geth. With batch requests enabled, multiple HTTP requests can be packaged into one single request and sent to the server. Server process this bulk request and returns a bulk result. All of these are done in a single round trip.

This feature can be useful for reducing the load on the server and improving the performance to a certain extent.

How to implement

To implement an HTTP batch request, just send a request with a payload containing multiple request objects in an array like below:

[
    {"jsonrpc":"2.0","method":"web3_clientVersion","params":[],"id":1},
    {"jsonrpc":"2.0","method":"eth_blockNumber","params":[],"id":2},
    {"jsonrpc":"2.0","method":"eth_syncing","params":[],"id":3},
    {"jsonrpc":"2.0","method":"eth_chainId","params":[],"id":4}
]

The server sends back the results in one response. The results are arranged in the same order as the requests were received. For example:

[
    {
        "jsonrpc": "2.0",
        "id": 1,
        "result": "Geth/v1.10.26-stable-e5eb32ac/linux-amd64/go1.18.8"
    },
    {
        "jsonrpc": "2.0",
        "id": 2,
        "result": "0x10058f8"
    },
    {
        "jsonrpc": "2.0",
        "id": 3,
        "result": false
    },
    {
        "jsonrpc": "2.0",
        "id": 4,
        "result": "0x1"
    }
]

To run it in curl:

curl 'YOUR_CHAINSTACK_ENDPOINT' \
--header 'Content-Type: application/json' \
--data '[
    {"jsonrpc":"2.0","method":"web3_clientVersion","params":[],"id":1},
    {"jsonrpc":"2.0","method":"eth_blockNumber","params":[],"id":2},
    {"jsonrpc":"2.0","method":"eth_syncing","params":[],"id":3},
    {"jsonrpc":"2.0","method":"eth_chainId","params":[],"id":4}
]'

Popular Web3 libraries like web3.js and ethers.js support this feature too. Below is an example of getting ether's balance from multiple accounts at once using web3.js.

const Web3 = require("web3");
const NODE_URL = "YOUR_CHAINSTACK_ENDPOINT";
const web3 = new Web3(NODE_URL);

const addresses = [
  "0x1f9090aaE28b8a3dCeaDf281B0F12828e676c326",
  "0x2bB42C655DdDe64EE508a0Bf29f9D1c6150Bee5F",
];

async function getBalances() {

  const startTime = Date.now()

  // Create a batch request object
  const batch = new web3.BatchRequest();

  // Loop through each address and add a getBalance request to the batch
  addresses.forEach((address) => {
    const request = web3.eth.getBalance.request(address, (error, balance) => {
      if (error) {
        console.error(error);
      } else {
        var timeFromStart = Date.now() - startTime
        console.log(`${address} has a balance of ${Number(web3.utils.fromWei(balance, 'ether')).toFixed(3)} ETH retrieved in: ${timeFromStart / 1000} seconds.`);
      }
    });
    batch.add(request);
  });

  console.log("Batch request:");
  console.log(batch.requests);

  // Send the batch request
  batch.execute();
}

getBalances();

The getBalances function creates a new BatchRequest object using web3.BatchRequest().

The function then loops through each address in the addresses array and creates a new request to get the balance of that address using web3.eth.getBalance.request(). It adds each request to the batch using batch.add().

Finally, the function executes the batch request using batch.execute(). When executed, the requests in the batch are sent to the Ethereum network simultaneously, and the callback functions are executed when the responses are received.

Multicall contract

A multicall contract is a smart contract that takes in the function call objects as parameters and executes them together. A developer can use the multicall contract as a proxy to call other contracts on Ethereum.

The implementation of a multicall contract is, in fact, very simple: it leverages Solidity’s call function to broadcast contract calls. This is a sample implementation of multicall’s aggregate function:

function aggregate(Call[] memory calls) public returns (uint256 blockNumber, bytes[] memory returnData) {
        blockNumber = block.number;
        returnData = new bytes[](calls.length);
        for(uint256 i = 0; i < calls.length; i++) {
            (bool success, bytes memory ret) = calls[i].target.call(calls[i].callData);
            require(success);
            returnData[i] = ret;
        }
    }

In summary, this function takes an array of Call, calls each one, and returns an array of the results along with the block number in which the function was called. It is designed to be used as a general-purpose aggregator for calling other contracts on the Ethereum blockchain.

How to implement

Anyone can deploy their own multicall contract. In this article, we leverage MakerDAO’s multicall contract on the Ethereum mainnet; which is deployed at 0xeefBa1e63905eF1D7ACbA5a8513c70307C1cE441.

Below is an example calling the smart contract with MakerDAO’s helper library multicall.js; it essentially does the same thing as the previous example:

const multicall = require("@makerdao/multicall")

const config = {
    rpcUrl: "YOUR_CHAINSTACK_ENDPOINT",
    multicallAddress: "0xeefba1e63905ef1d7acba5a8513c70307c1ce441"
};

const addressArr = [
  "0x2B6ee955a98cE90114BeaF8762819d94C107aCc7",
  "0x2bB42C655DdDe64EE508a0Bf29f9D1c6150Bee5F"
];

async function main() {
  const startTime = Date.now();
  console.log("Started...");

  const calls = [];

  // Retrieve the Ether balance of each Ethereum address in addressArr using the multicall library.
  for (let i = 0; i < addressArr.length; i++) {
      const callObj = {
          call: [
              'getEthBalance(address)(uint256)',
              addressArr[i]
          ],
          returns: [
              [`ETH_BALANCE ${addressArr[i]}`, val => val / 10 ** 18]
          ]
      };
      calls.push(callObj);
  }

  const result = await multicall.aggregate(calls, config);
  console.log(result);

  const timeFromStart = Date.now() - startTime;
  console.log(`Result received in ${timeFromStart / 1000} seconds`);
}

main();

The main function iterates through each address in the addressArr array and creates a call object for each address. These call objects use the multicall library to retrieve the ether balance for each address.

Once all of the call objects have been created and pushed to the calls array, the multicall library's aggregate function is called with the array of call objects and the configuration object. This function aggregates the results of all of the calls into a single object, which is stored in the result variable.

Finally, the code logs the "result" to the console and calculates the time it took to receive the "result", which is also logged to the console.

You will need to install the multicall.js library to run this code.

Performance comparison

In this section, we compare the performance of 3 different approaches:

Sending multiple HTTP requests in parallel
Sending a batch HTTP request
Using a multicall contract

We will test based on two common use cases:

Getting account balance
Calling a smart contract

Getting account balance for 30 distinct addresses

The testing script for batch requests and multicall contract is already included in the previous sections. Below is the code for sending multiple HTTP requests in parallel:

const Web3 = require('web3');
const web3 = new Web3(new Web3.providers.HttpProvider('YOUR_CHAINSTACK_ENDPOINT'));

var addressArr = [
    "0x2B6ee955a98cE90114BeaF8762819d94C107aCc7",
    "0x2bB42C655DdDe64EE508a0Bf29f9D1c6150Bee5F"
]

async function main() {
    var startTime = Date.now()
    console.log("started")
    for (i = 0; i < addressArr.length; i++) {
        web3.eth.getBalance(addressArr[i]).then(function(result) {
            var timeFromStart = Date.now() - startTime
            console.log("Result received in:" + timeFromStart / 1000 + " seconds")
        })
    }
}

main();

Result

	Parallel single requests	Batch request	Multicall
Round 1	1.789	1.49	1.447
Round 2	1.896	1.159	1.54
Round 3	2.337	1.113	2.132
Round 4	2.942	1.224	1.609
Round 5	1.638	1.602	2.012

The test was conducted between a server in Europe and a client in Singapore. A total of 15 measurements were averaged, which shows the performance of batch request > Multicall > normal request.
Compared with sending single requests in parallel, batch request reduces 38% of the total request time, and multicall reduces 18% of the total request time.

Getting the owners of BAYC tokens

Below are the testing scripts using web3.js for making smart contract calls. The tests are based on an ERC-721 standard method ownerOf from BAYC’s smart contract.

Sending multiple HTTP requests in parallel:

const Web3 = require('web3');
const web3 = new Web3(new Web3.providers.HttpProvider('YOUR_CHAINSTACK_ENDPOINT'));

const abi = [{ "inputs": [{ "internalType": "uint256", "name": "tokenId", "type": "uint256" }], "name": "ownerOf", "outputs": [{ "internalType": "address", "name": "", "type": "address" }], "stateMutability": "view", "type": "function" }]
const contract = new web3.eth.Contract(abi, "0xBC4CA0EdA7647A8aB7C2061c2E118A18a936f13D");

async function main() {
    const startTime = Date.now()
    console.log("started")
    for (i = 0; i < 30; i++) {
        contract.methods.ownerOf(i).call().then(function(result) {
            console.log(result)
            var timeFromStart = Date.now() - startTime
            console.log("result received in: " + timeFromStart / 1000 + " seconds")
        })
    }
}

main();

Sending batch request:

const Web3 = require('web3');
const web3 = new Web3(new Web3.providers.HttpProvider("YOUR_CHAINSTACK_ENDPOINT"));

var abi = [{ "inputs": [{ "internalType": "uint256", "name": "tokenId", "type": "uint256" }], "name": "ownerOf", "outputs": [{ "internalType": "address", "name": "", "type": "address" }], "stateMutability": "view", "type": "function" }]
var contract = new web3.eth.Contract(abi, "0xBC4CA0EdA7647A8aB7C2061c2E118A18a936f13D");

async function main() {
    var startTime = Date.now()
    var batch = new web3.BatchRequest();
    console.log("started")

    for (i = 0; i < 30; i++) {
        batch.add(contract.methods.ownerOf(i).call.request(function(err, result) {
            console.log(result)
            var timeFromStart = Date.now() - startTime
            console.log("Result received in: " + timeFromStart / 1000 + " seconds")
        }))
    }
    batch.execute();
}

main();

Multicall contract:

const multicall = require("@makerdao/multicall")
const config = {
    rpcUrl: "YOUR_CHAINSTACK_ENDPOINT",
    multicallAddress: "0xeefba1e63905ef1d7acba5a8513c70307c1ce441"
};

async function main() {
    var startTime = Date.now()
    console.log("started")
    var calls = []

    for (i = 0; i < 30; i++) {
        var callObj = {
            target: "0xBC4CA0EdA7647A8aB7C2061c2E118A18a936f13D",
            call: ['ownerOf(uint256)(address)', i],
            returns: [
                ['OWNER_ADDR ' + i]
            ]
        }
        calls.push(callObj)
    }

    const result = await multicall.aggregate(calls, config);
    console.log(result.results);

    var timeFromStart = Date.now() - startTime
    console.log("Result received in: " + timeFromStart / 1000 + " seconds")
}

main();

Result

	Parallel single requests	Batch request	Multicall
Round 1	1.693	1.931	1.878
Round 2	1.717	1.592	1.195
Round 3	1.712	1.617	2.183
Round 4	2.103	1.589	1.3
Round 5	2.785	1.416	1.429

The same test was conducted for read contract calls. The result shows that both batch requests and multicall contracts save around 20% of total request time compared with sending single requests.

Common questions

Question 1. If I package 100 requests into a single batch request, does that count as 1 request or 100 requests on Chainstack?
Answer. As an RPC provider, Chainstack counts “request” as RPC calls. After a server receives an HTTP batch request, it “unpacks” the request and processes the calls separately. So from the server’s point of view, 1 batch request of 100 calls consumes 100 requests instead of 1.

📘
Check the Understanding your request consumption page on Chainstack support docs for more details.

Question 2. If I package 100 calls into a single multicall request, does that count as 1 request or 100 requests?
Answer. In this case, even though it is a very heavy call, it counts as a single request.

Question 3. Is there any hard limit for the number of calls to multicall contracts?
Answer. The BAYC testing script stops working with 1,040 calls.

Which approach works better for me

Even though tests show that batch request and multicall contract improves performance significantly, they do have their own limitations.

Requests in a batch request are executed in order, which means if a new block is received during execution, the subsequent results are likely to be different.

If you want to use a multicall contract, you should probably deploy your own contract for production just to ensure its availability.

Both batch request and multicall contract return multiple results in a single response. Both of them require much more computational resources. They can easily trigger “request timeout” errors and “response size too big” errors, which makes them not suitable for complex calls.

📘
See also

HTTP batch request VS multicall contract

How to store your Web3 DApp secrets: Guide to environment variables

Goerli ➡️ Sepolia transition

HTTP batch request VS multicall contract