Topic: TXUO - Bitdust questions (Read 756 times)

Its just rolling around in my head...  44gb not scary with todays ram (of course ALL coins would never all be distributed to the smallest unit).  But suppose 10% or roughly 4gb would be needed.  Thats within reach of the average enthusiast.  This is what the smallest output (not necessarily the smallest transfer unit) would have to be when sending to a *new* address.  You could still enjoy yourself and send .000001 to a preexisting address.  My understanding is that sending coin to an address already containing the minimum unspent output  does not create a new TXUO.  Or does it?  (please let me know protocol whizes!) Then in that case .01 would be the smallest unit you could send if we wanted revise the protocol to guarantee that the blockchain could be run on the hobbyist computer.

Remember, part of the promise of bitcoin was that any one could 'be their own bank.'  It serves no one if bitcoin becomes popular, UTXO blows up and causes every on to move to a few online wallet services and then the sleeping giants in Washington and Brussels wake up and yank the domains of blockchain.info and their ilk.

A distributed way of doing small coinage might be doable.  A blockchain is an inefficient way to manage a database of who paid who, a central server is better.  If a system of central servers, with a low barrier to entry, existed, say several hundred spread around in different jurisdictions, could just use the blockchain to periodically settle net payments, and people just transacted between central servers their bitdust, bitcoin could still work very nicely.

It depends what you think "Bitcoin" is.  Its not JUST the blockchain (or at least doesn't have to).  A distributed, peer to peer, Open Private Transaction Processing System could be overlayed as an extension to the protocol.  Some don't like that idea because it gives them visions of Visa and Paypal, or even MyBitcoin.com scandals.  It doesn't have to be that way either you know, if its designed right, it could be trusted enough (or at least diversify your balance through a group of OPTPSs) and have low enough barriers to entry that all one would need is to run the Open Source OPTPSs host software, publish a public key, have those endorsed by competing OPTPSs, and keep up with the blockchain.  If the blockchain starts to become too costly, or is limited to only larger txs b/c a minimum TXUO value is established, then people will do most day to day stuff thru OPTPSs.  The purists and tin foil hat wearing folks can stick to the blockchain and their paper privkeys stored next their bugout bag, AR-15 and MREs.  Then can pay the high fees.

Last i checked, in the old days, banks didn't need to settle $5 balance differences.  The waited till a competing banks checks piled up then wired them to arrange a settlement in physical gold.  It was expensive, slow an dangerous to move a large amount.  BTC distributed blockchain is the modern equivalent of that.  

Now some say all this can be overcome by optimizing the blockchain, pruning, clusterfucking IDK.  But if it can't, that will likely express itself as a rising tx fee.  When that happens i'd like to leave it to an open source tps system that we have in the bullpen running on the testnet.  otherwise Bitpay, Mtgox, etc will step in and it will be paypal all over again.  Not to mention that OPTPSs can give eachother instant confirms, with prompt settlement via the blockchain.

You want people to be only able to spend ~$1 (at current ex rates), and likely $2-$10 later on?

Thanks, then if the minimum currency unit was .01 the max unspent outputs would only be 44 gb

I'm trying to put together an idea for dealing with blockchain gerth.   Specifically, as I understand the most pressing aspect of blockchain growth is the ballooning size of  "unspent outputs" (TXUO?).  That is, if usage patterns continue to scale, with the majority of tx's coming in as bit dust, the first wall we will hit is a TXUO that gets too large to be held in "fast memory" for the average user.  The concern is that everyone would be forced to a cloud wallet service, which currently is populated by only a handful of players, and could serve to force many needed miners out of verifying the blockchain.  This also creates a choke point at this point in BTC development that regulation happy governments can pinch.

Obviously, the size of the actual blocks is a concern itself, in terms of bandwidth, disk storage costs and possibly compute power to verify all tx's inside each block.  But those issues as I understand take a back seat to how large TXUO is getting, and the need to hold it in "fast" memory.  (i know that there is also a hard 1 mb limit on blocks, but lets set that aside as a limiting factor for now).

So my first question:
1) Is the expanding TXUO going to be the first bottleneck the network will hit?

If so,

2) What is the current size of TXUO and is it held in main memory by miners/satoshi client?  (if not, at what size does it necessitate being held as such).

3)  How much memory does a single TXUO take up?  And as a following question: assuming all BTC were issued today, and ALL 21 mil were broken up into addresses of 1 satoshi, how big could TXUO theoretically get?

I have an idea for extending/revising the BTC protocol to deal with this, but i first need to know if this really is the core problem.  If we hit a blocksize, or bandwidth limit first, then thats what i need to know as the problem.  Again, the problem is defined as one of these limit factors causing miners to drop out, or mainline users being forced into a handful of "regulatable" wallet services.

So, what IS the problem we need to solve here?