A transfer request expands through the graph as usual, never rebroadcast if seen before. There are two outcomes, a double spend collides, and a revoke is released for both, which expands and destroys both spend partitions.. All transfers timeout, to revoke, if they are not concluded.
The count of miners is known, when each miner cannot find an output path, it releases the clear. The clear accumulates back to the source with a miner count equal to total known miners. Works great for networks of 20 thousand nodes, Completing in nlog n time because it i inherently multi-processed, a graph algorithm.
The miner network can be distributed about, but traders will ten to cluster over the high speed network. They will position their bots right on top of the data center net. The transactions should clear in milliseconds, as fast a context switch time in the processor. Traders have their miner code under some control, prequaled and code verified on demand. Kind of like gun control.
The system works best when each miner can issue public keys to neighbors, and hold a private key. The protocol is simple can be done in thread switching time as long as the processor can protect the thread private key.
Put the miners in a barely threaded system, run through each transaction request with array speed, keep the environment secure. But swap net is web based, and in its most secure state relies on some bonded carrier of keys. They need the armored box, a known device with hardware validation,it is counterfeit proofed. Traders buy a spot on the device, and only it generates public keyes and this armored box protects private keys, and can generate them. It plugs into the net work. There is enough here to ship public keys around, secretly, and validate them with the private key, in the box, on a mass basis in a high speed network,
Multiple liquidity graphs
From the traders point of view, the swap net group could be running multiple liquidity graphs. One for bitcoin, fiat, or anything. The life of pure liquidity is finite, it must mark to market in the exchange and returned to the external ledger.
Traders organized, directed graphs
In the data center, traders can own secure space, and swap net is the local high speed hardware, so we context switch and swap across the local optical net. Some traders get a lot of traffic, Let them spread the work load, split the request and pass them down the directed graph. Then each node can check the smaller batch and send revokes or clear back up the graph. You are assured of NlogN time, I think. Traffic management can keep the minimum connected graph, for distant data centers, each can batch up the largest set and ship it to the other.
Graphs constructed to minimize congestion, connections verify with public key neighbors. Counts known and finite,only 100% agreement allowed.
Where is the coin?
The fair transaction checksummes the collected signatures, with the transaction. It ids directly verifiable by any miner, identifies the holding trader account. Once the trader swaps its encoded trace, the new trader gets the new 100% trace, and the previous trace deleted. That is nearly pure liquid form for any established currency.
Congestion
The graph is bidirectional, miner node store state. So, each miner checks double spends and issues conclusions. If work piles up, it ships unconfirmed transactions down stream. Clear or revoke will cover the graph and accumulate back to the source,
Add accounting rules to the nodes
How long should a miner wait for a revoke? After a timeout and no return revoke a miner can erase its partial verification, forget the transactions. Miners can execute regulation requirements, exchange congestion information such as to rearrange the graph.
Delegation
Not all traders need run a mining process, they can thumbprint PoA to a trusted miner. We het large dispersed clusters.
Global Swap times
If the network is truly unqueued, then a global swap net could clear one or two a second for the trader. But if the liquidity nets mostly trade inside one block of ice, economies of scale, global rate in the forty or fifty trades per second for the trader.
Where is the coin?
The fair transaction checksummes the collected signatures, with the transaction. It ids directly verifiable by any miner, identifies the holding trader account. Once the trader swaps its encoded trace, the new trader gets the new 100% trace, and the previous trace deleted. That is nearly pure liquid form for any established currency.
Congestion
The graph is bidirectional, miner node store state. So, each miner checks double spends and issues conclusions. If work piles up, it ships unconfirmed transactions down stream. Clear or revoke will cover the graph and accumulate back to the source,
Add accounting rules to the nodes
How long should a miner wait for a revoke? After a timeout and no return revoke a miner can erase its partial verification, forget the transactions. Miners can execute regulation requirements, exchange congestion information such as to rearrange the graph.
Delegation
Not all traders need run a mining process, they can thumbprint PoA to a trusted miner. We het large dispersed clusters.
Global Swap times
If the network is truly unqueued, then a global swap net could clear one or two a second for the trader. But if the liquidity nets mostly trade inside one block of ice, economies of scale, global rate in the forty or fifty trades per second for the trader.
No comments:
Post a Comment