EOS.IO Stack Exchange is a question and answer site for users and developers of distributed applications built on top of the blockchain operating system, EOS.IO. It only takes a minute to sign up.
Sign up to join this community
I am running so called SuperHero testnet node on my computer, and I've noticed that according to logs, blocks are produced in series, usually by 12 in one series produced by one block producer:
Can we expect exactly the same behavior on mainnet or this is testnet specific behavior?
Assuming, this will also work like that on mainnet, how many new blocks should confirm a block A, so a typical application could be sure, that a block A will not be reversed?
In the current specification, each of the 21 Block Producers will produce 6 seconds worth of blocks (i.e. 12 blocks) before handing off to the next producer in the production order.
Irreversibility is not determined by number of blocks, but by the number of producers who build on blocks. The protocol will tell you when a block becomes irreversible.
Once BFT messages are implemented all other producers will "ack" every block proposed and give irreversibility as fast as 1 second under good network conditions. Absent BFT messages, the DPOS LIB rules should make the block irreversible after about 1.5 minutes.
The most disruptive part of consensus is the handoff between producers, a producer should never fork their own blocks. By only switching once every 6 seconds users get fast confirmations with relatively few missed blocks. A well optimized network may not miss any blocks during the handoff period.
6 seconds was chosen based upon the "maximum downtime" if a producer goes off line. This matches Steem & BitShares where a single missed block creates 6 seconds without any confirmation.
In the current specification, each of the 21 Block Producers will produce 6 seconds worth of blocks (i.e. 12 blocks) before handing off to the next producer in the production order. (see bytemaster's answer below for more details)
