Thread by @LN_Master_Hub

#bitcoin #LightningNetwork Multi-party channels (NOT channel factories) offers massive benefits over standard 2 party channels.
Fees to open are minimal.
Finding path becomes MUCH MUCH easier.
From a routing perspective, if you have 1 BTC in a MP-channel, it can be routed to

https://twitter.com/LN_Master_Hub/status/1116178075522174976

any other member of the MP channel. Paths becomes drastically shorter, and cheaper, while simultaneously increasing liquidity massively. It has a multiplicative factor on all liquidity.
Channel factories = BTC split amount x number of channels
MP-channel = a channel for the

full amount to every single other member. So, if I join a 1001 member MP channel with 1 bitcoin, it'll show up a 1,000 BTC of liquidity. The channel graph would show me having 1,000 channels with 1 BTC each, ready and willing. (PLUS the amount each other person has to me)

So, for routing purposes, say 1000 people each joined with 1 BTC (1000 BTC total).
This would show up as 1000 people each having 1 BTC with 999 other people creating 999,000 BTC of liquidity with just 1,000.

Then, lets say each of those 1,000 people opened a 2nd channel to another 1,000 person MP-channel, assuming no overlap, any person could reach over 1,000,000 people with just two hops.

Also, due to the TX size savings, a couple hundred channels could be opened per block with 1000 ppl each.
It'd be possible to onboard 200,000 people to LN per block through automated MP-channel creation services.

MP-channels solves several problems at once:
High-fee scenario
Liquidity
# of hops to get to destination becomes dramatically shorter for large populations... 4 hops from anywhere to anywhere with 10,000,000,000 nodes.
"Centralization" of nodes AKA MasterHub scenario

MP-channels do come with trade-offs, as well as some solvable problems.
Tradeoff:
Lose some privacy, as everyone in the MP knows you balance and when it changes.

Problem:
Current smart contracts requires signatures of all members.
This will have to be altered in a manner

that allows for having multiple "parts" with in the 'contract' and you know need to sign when your own part changes, and you only *have* to store your own, but can store all if you choose. (Looks a lot like full node VS neutrino, and requires trust if you're not storing all)

This would allow any number of member nodes to be offline without affecting all other member's ability to transact.

When Multiparty channels @alexbosworth @starkness @adam3us ?

https://twitter.com/LN_Master_Hub/status/1116184260006670337

As far as I know... There isn't a single person doing even the slightest bit of research into MP-channels. Everyone saw what I listed as a 'solvable problem' and flat out didn't bother trying to solve it. It's solvable, but complex and difficult to solve.

It would take a complete overhaul to how 'smart contracts' are thought of conceptually, and coded.
Right now, in a conceptual MP-channel, everyone has to sign every LN transaction, even ones they are not a part of. This requires 100% of members to be online 100% of the time.

This cripples it's usefulness.
One would have to design a method where each node controls a 'part', aka their balance, and they only need to sign when their balance changes. Then broadcast their new balance to everyone else. Exactly how bitcoin works onchain.

Nodes could go offline, then when they come back online, ask other members for any broadcasts they missed, and verify with others.
Exactly how bitcoin works onchain. ;)

Essentially, each MP-channel would function like a side-chain, instead of offchain.

But with the added benefit of being a "LN channel", they could be daisy-chained together, just like 2-party channels.

And the onboarding rate goes from 3/s to 500/s, leaving room for other transactions as well.
With MP-channels, you could reasonably onboard 40 million people per day for 5 cents each, even if fees are $50, all with 1 MB blocks.

An additional benefit: Pathfinding time could be dramatically reduced.
Instead of checking each of the 999,000 different potential connections in 1000 node MP channel, the aglo could check the blob as a whole, and see if blob A connects to blob B which contains the destination.

Once it finds the correct path, blob to blob, it then goes deeper to find the cheapest connections within said blobs.

With billions of people, and possibly trillions of channels, this reduces the pathfinding overhead by orders of magnitude, while essentially putting the entire

planet at most 4 blob-hops from any other person on the planet.

This effectively destroys the concept of "routing nodes", as everyone would be very very closely connected to everyone else.
Though, the centralization would be extreme... everyone on the planet centralized in one place. ;)

AMP becomes pretty pointless as well, due the exponential effect on liquidity that MP channels have, as was as the excessively reduced path.

#Bitcoin #BTC and #LightningNetwork have so many 'orders of magnitude' efficiency upgrades just waiting to be discovered/implemented... The world will not be the same in 10 years as it is today.
Each upgrade compounding the next... billions of times better than it is today.