Migrating to RocksDB will make running full nodes a lot cheaper which will make RC's cheaper, too, which is a good thing because it's not the high cost of transacting on the blockchain that makes STEEM valuable but all the traffic on it. The lower the transaction costs the better. That achievable by running efficient RPC nodes. What we need is masses of people using Steem DApps. Realistically, we can expect valuation in the low tens of billions when we have hundreds of millions of users - or perhaps tens of millions of users if enough people are willing to pay for the advantages of decentralization. We are extremely far from that kind of numbers but making transactions drastically cheaper is the way to get there.
You are viewing a single comment's thread from:
That’s been my impression as well, and my main hope for scalability. For all the good points of Steem scarcity listed in this article, the big problem is at current conditions we’re capped at a relatively small number of normally functional user accounts.
In a recent State of Steem forum on Discord, I was intrigued most by some of the discussion from @blocktrades. Apparently their team was the original subcontracted developers of initial RocksDB implementations (the user account history module I think it was?) He also confirmed that upcoming node efficiency and cost reductions could bring down resource costs, because they would make more resources available for everyone.
He even went on to describe a future where it could feasibly be possible to have simple node install packages for consumer hardware that even a layman could use. I envision it like most proof of stake wallets that stay open & running to secure the network while staking rewards. We could see Steem clients that basically act as nodes and help generate the network resources the end user is consuming.
All a bit pie in the sky, and I’m very much the “layman” I mention above... but I feel there are smart folks working on the technicals, and many outside of Steemit Inc. capable of taking up the tasks if need be. That keeps my interest and optimism alive... perhaps foolishly!
Yes, that’s correct. But let’s make sure not to confuse ‘Resource costs’ (RC) and ‘Resource costs’ (operational infrastructure costs).
The latter being how much it costs to host a node on a server. The specific hardware requirements for such. We’ve known for decades that they were ridiculous for such a small community as Steem currently is.
The former being the”new bandwidth” matrix introduced with HF20. And then hotfixed by applying a
10xcalculation.My point is that unless you can create proven value, nobody is going to pay high “entry fees” to be capable to operate.
RocksDB is challenged in mseveral points already. It’s a long overdue focus but at the same time several have highlighted already that it does away with other benefits inherent to graphene tech. RocksDB should also not change the economics of the RC model. That would be the biggest clusterfrak yet by STINC if.
It’s just forgotten nowadays that we operate at 10x RC.
At the same time, 5 million users [at current cost] is 150m STEEM burnt.
PS: RCs will operate from “delegated pools”, so more likely than not be available to dApps. AFAIK there is no user-to-user RC delegation model yet. Also... right now all that is vaporware.
We would have received the initial SMT testnet December 2017 (as per Ned in SMT Telegram). We will eventually receive “something someday” but while before nobody knew what, now we know even less.