who could tell me why the IC update query works so slowly(typically takes more than 4s)?
Is there has some possible method to improve IC’s update query time? (let’s say improving it to 0.5s are great for application running on ic…)
Update calls go through full consensus (in the subnet) and take a few seconds. This is because ~15-30 nodes need to all agree on an action and only then canister state is able to be modified.
Query calls are less secure as they only run on a single node (bad actor risk) but they do run in a few hundred ms. We can also provide cryptographic proofs/witnesses for queries allowing a client to verify the data is true, mitigating bad actor attacks.
3 Likes