BitShares Forum

Main => Technical Support => Topic started by: hamiltino on October 01, 2015, 12:47:18 am

Title: What is the block time of bitshares 2.0 on release?
Post by: hamiltino on October 01, 2015, 12:47:18 am
Is it 1 second? Or will that be down the track?
Title: Re: What is the block time of bitshares 2.0 on release?
Post by: Ander on October 01, 2015, 12:48:31 am
3 seconds initially, 1 second later on.
Title: Re: What is the block time of bitshares 2.0 on release?
Post by: Troglodactyl on October 01, 2015, 02:10:23 am
3 seconds initially, 1 second later on.

If people even want 1 second blocks, really.  It will be voted on, but personally I don't see much benefit of 1 second over 3 second.  Seems like it's just for bragging rights and makes chain bloat.
Title: Re: What is the block time of bitshares 2.0 on release?
Post by: 38PTSWarrior on October 01, 2015, 03:05:30 am
They are all paying with their no-touch cards now and this goes faster than 3 seconds. Plus 5 % for 1 sec!
Title: Re: What is the block time of bitshares 2.0 on release?
Post by: Erlich Bachman on October 01, 2015, 05:09:46 am
3 seconds initially, 1 second later on.

If people even want 1 second blocks, really.  It will be voted on, but personally I don't see much benefit of 1 second over 3 second.  Seems like it's just for bragging rights and makes chain bloat.

For real. 10 second blocks seem a bit clunky when trading at times, but 3-5 second blocks are close enough to real time where you start experiencing diminishing returns even for those like myself who are queing up several trades back to back.  Make this parameter a voting parameter and let the masses decide what they want but you can still brag that the capability is there if enough traders want to buy stake and vote 1second block times into office.
Title: Re: What is the block time of bitshares 2.0 on release?
Post by: mint chocolate chip on October 01, 2015, 05:15:56 am
Quote
With Delegated Proof of Stake, the BitShares network can confirm transactions in an average of just 1 second, limited only by the speed of light.
https://bitshares.org/technology/industrial-performance-and-scalability/
Title: Re: What is the block time of bitshares 2.0 on release?
Post by: Troglodactyl on October 01, 2015, 12:32:07 pm
3 seconds initially, 1 second later on.

If people even want 1 second blocks, really.  It will be voted on, but personally I don't see much benefit of 1 second over 3 second.  Seems like it's just for bragging rights and makes chain bloat.

For real. 10 second blocks seem a bit clunky when trading at times, but 3-5 second blocks are close enough to real time where you start experiencing diminishing returns even for those like myself who are queing up several trades back to back.  Make this parameter a voting parameter and let the masses decide what they want but you can still brag that the capability is there if enough traders want to buy stake and vote 1second block times into office.

It's definitely going to be subject to vote.  That's already decided, but it's still worth debating the costs and benefits here in order to educate the voters.  There's definitely a point of diminishing returns. 

1 second blocks just have very low tolerance for network slowdowns and increase the overhead costs in terms of both bandwidth and storage space.

Quote
With Delegated Proof of Stake, the BitShares network can confirm transactions in an average of just 1 second, limited only by the speed of light.
https://bitshares.org/technology/industrial-performance-and-scalability/


...Which implies a block time of 2 seconds.
Title: Re: What is the block time of bitshares 2.0 on release?
Post by: bytemaster on October 01, 2015, 12:43:07 pm
As monster said in another thread, to be absolutely certain that a witness cannot double spend against you you must wait for 2 blocks which means that 3 seconds becomes 6 and 1 becomes 2. 

There is a fixed cost per block that the network must pay that is independent of transaction volume.   At 1 second blocks it becomes 8 MB per day of empty blocks.   Having 2 second blocks would save 1.5 GB per year.

That overhead is insignificant for full nodes and doesn't impact 99% of users of light nodes. 
Title: Re: What is the block time of bitshares 2.0 on release?
Post by: btswildpig on October 01, 2015, 01:04:25 pm
As monster said in another thread, to be absolutely certain that a witness cannot double spend against you you must wait for 2 blocks which means that 3 seconds becomes 6 and 1 becomes 2. 

There is a fixed cost per block that the network must pay that is independent of transaction volume.   At 1 second blocks it becomes 8 MB per day of empty blocks.   Having 2 second blocks would save 1.5 GB per year.

That overhead is insignificant for full nodes and doesn't impact 99% of users of light nodes.

One dumb question :

So regarding decentralized exchange , does this mean a deal can be reversed or a UIA can be sold to multiple people  at the same time by a witness ?
Title: Re: What is the block time of bitshares 2.0 on release?
Post by: bytemaster on October 01, 2015, 01:11:26 pm
As monster said in another thread, to be absolutely certain that a witness cannot double spend against you you must wait for 2 blocks which means that 3 seconds becomes 6 and 1 becomes 2. 

There is a fixed cost per block that the network must pay that is independent of transaction volume.   At 1 second blocks it becomes 8 MB per day of empty blocks.   Having 2 second blocks would save 1.5 GB per year.

That overhead is insignificant for full nodes and doesn't impact 99% of users of light nodes.

One dumb question :

So regarding decentralized exchange , does this mean a deal can be reversed or a UIA can be sold to multiple people  at the same time by a witness ?

It means that if you do something irreversible off-chain in response to something you see happen in the head block that the witness could "double sign" and create another block that includes different transactions.

Any witness that would double-sign would get voted out immediately (and their peer witnesses should be configured to ignore that witnesses blocks).   

In practice it is a purely theoretical attack and will not happen because the witnesses will be public (non-anonymous) and they would not make much money.   

To pull of the attack also requires that the attacker identify YOUR node and intentionally send you a block.   If you are not directly connected to the attacker and the attacker does not know where your computer is on the network then the attack becomes technically impractical.   

In other words, only the truly paranoid would worry about it.  You are far more likely to get struck by lightning.