Quote from: MalReynolds on October 01, 2014, 07:22:32 PM

Quote from: starik69 on October 01, 2014, 05:11:25 PM

Quote from: Litoshi on October 01, 2014, 04:35:48 PM

The forks were due to the short 60 second block timing, and that the node was communicating with every other node in that minute. This worked well with 50 nodes on testnet, but did not work when scaled up. The Proof of Time algo was a great idea, to reward loyal forgers, but it is a bottleneck at present.

OMG!

They really thought that such simple algo will work on network with >1000 nodes? :doublefacepalm: Roll Eyes

It's obviously not scalable to have "everybody talking to everybody during all blocktimes".

For PoT to be implemented, the total pool of nodes is going to have to be divided into a number of subpools where only nodes in one subpool have the "possible right" to forge during a given blocktime, only one node in this subpool actually DOES forge during this blocktime, only one choosen subpool talks "everybody to everybody" during this blocktime, and all other subpools must STFU during this blocktime. The max subpool size is dependent on what will work in the real world, and the "devil in the details" is how nodes are added and subtracted to an appropriate subpool as they come online or go offline.

Not an impossible problem, but tricky, and it would avoid both the PoW ASIC arms race of Bitcoin and the PoS hoarding problem of NXT to set a new and viable 2.0 standard methodology. I'm hodling.

Hey Mal,

It's an interesting concept and similar to one I was discussing with someone as a possible solution. After we get the main net back up and running and get a chance to provide some details, I'd be interested in discussing this a little bit more in detail. We may try to setup a community chat once it calms down a little to discuss some of the options.