Frontend vs Backend RPC call performance

Questions
1

Hello,
I’m having trouble using multicall contract calls, seems like the same call including 7 eth_calls (I’m using the same data for benchmarking) takes me:

  • 800 ms using multicall with Infura
  • 200 ms using batch call with Infura
  • 200 ms using multicall with super random competitor RPC
  • 200 ms using batch call with super random competitor RPC

Could it be something with my config ?

at the moment I’m using axios without any particular configuration.

Regards

1 Like
2

Apparently the difference in fact comes from the source of the call. When I call the RPC from the frontend the response times are reasonable, but when called from the backend (nextjs API endpoint) the times are 4x longer. The time measured - is only the time required to perform POST call with axios.

  const multicallResponse = await axios.post<string, AxiosResponse<{ result: string }>>(
    provider.connection.url,
    callBody,
    config,
  )

What can be done to improve the response times? Does Infura require any particular headers or other call settings?

image

3

Hey, is your frontend and backend running on the same machine or location ?

4

Yes and yes.
I’ve tested it with nextjs locally and on AWS with same results.