Same algorithm I was using for shared memory space across distributed memory managers.
Any node issues a transaction, and it expands th consistency net until a transaction wins. The nodes are secure code, in this example. There is a cancel within timeout. All ledgers nodes keep the same transaction model, but differ in ledger type. Inside a liquidity net, this makes for great multi level trades, a three or four sided swap among pits and ledgers. Make it the etthernet of escrow routers. Inside the net, trusted managers have a great tool to subdivide the flow into pits, ledger and vaults (hot wallet).
Simple algorithm, and it allows for conditional trades, meaning the request is a contract, a finite graph not a scalar. I worked this problem a couple of times back in the day, it works. Create a little script to describe contracts as finite spanning tree with valid exits. Then you have a provable map from the written, human judge to the generic contract.
The default.
In the hand held, your portfolio adjustments are all routers, or router requests to trusted miners. When available, some set of contracts will execute, take a step up to the finite limit. So we an conceive of a variety of transactions over distributed nodes, if we ascertain small finite step spanning tree. Simple contracts,, timeout, inside Secure ID, great stuff, seriously.
No comments:
Post a Comment