QBFT Block time < 1 seconds

[pangtongs]

Rationale

Why should this feature exist?
I am using my application in a real production environment that requires latency to be as quick as possible and 1 second block time is not satisfactory enough for this implementation.

What are the use-cases?
To increase the latency of my application and not letting client of my application to have to wait too long before a transaction finish

Implementation

Do you have ideas regarding the implementation of this feature?
no

Are you willing to implement this feature?
no

[mipicdev]

Hi all,

I was just about to raise a similar issue, but it seems you're one step ahead :-)

In our test setup, we're operating with four validators and four Tessera nodes, each running in a separate Docker container. After tweaking the genesis.json, carrying out a number of deployments, and interacting with a smart contract, I've observed that transaction times typically range between 1 to 3 seconds.

While this is not excessively long, I believe there might be room for improvement. I'm particularly interested in exploring ways to reduce these transaction times further. Would anyone have any insights or recommendations on how we might accomplish this?

Looking forward to hearing your thoughts.

Best,
mipicdev

[Kaikj]

What is being brought up here is quite aptly summarized by #226 (comment)

I don't think the concerns from that issue has been fully addressed yet. @hmoniz or any other contributor, care to chime in on this topic? What are some of the next actions that can be taken to bring this conversation forwards?

[hmoniz]

Hi. I have not been directly involved in IBFT/QBFT since 2021. I designed the protocol and contributed to its first implementation, but I have not followed its direction since then. My opinion still stands though. While such a change is likely to require a significant engineering effort, there's no fundamental reason why it can't be done.