Bitcoin Forum
July 08, 2024, 03:54:12 AM *
News: Latest Bitcoin Core release: 27.0 [Torrent]
 
   Home   Help Search Login Register More  
Pages: [1]
  Print  
Author Topic: LN intermediary channels  (Read 1835 times)
jonald_fyookball (OP)
Legendary
*
Offline Offline

Activity: 1302
Merit: 1004


Core dev leaves me neg feedback #abuse #political


View Profile
April 08, 2017, 01:05:31 AM
 #1

So lets "A" wants to send a payment to "G"
and there are 5 bidirectional payment channels:

A->B
B->C
C->D
E->F
F->G

Questions:

1)  Can "G" get his money settled
on the main blockchain by closing out his channel
with F, even if the other 4 channels remain open
indefinitely?

2) What happens to A's money that
she sent to G if she never closes
her channel with B?

3) What are the implications for B,C,D,E,F ?





achow101
Moderator
Legendary
*
expert
Offline Offline

Activity: 3444
Merit: 6748


Just writing some code


View Profile WWW
April 08, 2017, 01:48:15 AM
 #2

1)  Can "G" get his money settled
on the main blockchain by closing out his channel
with F, even if the other 4 channels remain open
indefinitely?
Yes.

2) What happens to A's money that
she sent to G if she never closes
her channel with B?
G received an amount of money that A owed to him. A's money technically went to be.

3) What are the implications for B,C,D,E,F ?
Nothing.

stdset
Hero Member
*****
Offline Offline

Activity: 572
Merit: 506



View Profile
April 08, 2017, 06:27:38 AM
 #3

2) What happens to A's money that
she sent to G if she never closes
her channel with B?
They remain in A<->B channel. In LN funds belonging to a channel never leave that channel.

Example of Initial state of your piece of LN, channels are enclosed in parentheses:
A(1:0)B(3:2)C(4:1)D(1:2)E(5:6)F(2:0)G

After A pays 1 BTC to F (intermediaries don't charge fees):
A(0:1)B(2:3)C(3:2)D(0:3)E(4:7)F(1:1)G
Each channel contains the same amount of BTC as before. Note, that A can't pay F again through this route, because A's balance is 0, and D's balance (in A->G direction) is also 0.
If G closes his channel with F, they will receive onchain their shares of their channel and that's it.

Carlton Banks
Legendary
*
Offline Offline

Activity: 3430
Merit: 3074



View Profile
April 08, 2017, 10:32:21 AM
 #4

she sent to G if she

FYI jonald


Eazy-E (your avatar) was not much of a feminist. He was a rapper in NWA. Just in case you need to get your persona straight.

Vires in numeris
jonald_fyookball (OP)
Legendary
*
Offline Offline

Activity: 1302
Merit: 1004


Core dev leaves me neg feedback #abuse #political


View Profile
April 08, 2017, 02:17:59 PM
 #5

Thanks for the responses. lol Carlton.




Each channel contains the same amount of BTC as before. Note, that A can't pay F again through this route, because A's balance is 0, and D's balance (in A->G direction) is also 0.
If G closes his channel with F, they will receive onchain their shares of their channel and that's it.

so, this (bolded) is what is troubling me about LN... It seems that people would either have to keep re-opening channels or
the routing would have to keep going through bigger and bigger channels (higher balances), both of which seem
to defeat the purpose.

Has anyone done any study on how LN would function as a mesh network as opposed to more of a hub and spoke geometry?

Carlton Banks
Legendary
*
Offline Offline

Activity: 3430
Merit: 3074



View Profile
April 08, 2017, 03:01:36 PM
 #6

Has anyone done any study on how LN would function as a mesh network as opposed to more of a hub and spoke geometry?

It's a hybrid, it works as either peer to peer or as hub and spoke, depending on the way channels naturally form into an overall network. It's no different to TCP/IP in terms of network topographical properties. This topography and state (that you chose, remember) doesn't lend itself well to A paying F inherently, because of it's specific state and topography, and nothing to do with any fundamentals of Lightning. It's a jonald_fyookball fail, not a Lightning fail.

Have you really been objecting to LN frequently and consistently "because centralisation" in ignorance of this fact jonald?

Vires in numeris
jonald_fyookball (OP)
Legendary
*
Offline Offline

Activity: 1302
Merit: 1004


Core dev leaves me neg feedback #abuse #political


View Profile
April 08, 2017, 04:23:11 PM
Last edit: April 08, 2017, 04:38:19 PM by jonald_fyookball
 #7

Has anyone done any study on how LN would function as a mesh network as opposed to more of a hub and spoke geometry?

It's a hybrid, it works as either peer to peer or as hub and spoke, depending on the way channels naturally form into an overall network. It's no different to TCP/IP in terms of network topographical properties. This topography and state (that you chose, remember) doesn't lend itself well to A paying F inherently, because of it's specific state and topography, and nothing to do with any fundamentals of Lightning. It's a jonald_fyookball fail, not a Lightning fail.

Have you really been objecting to LN frequently and consistently "because centralisation" in ignorance of this fact jonald?

I finally respond to one of your posts after a year of ignore, and already the personal insults?

I don't recall 'objecting' to LN; obviously, the more decentralized, the better... and I am trying to determine the extent to which that will be possible.  I didn't intend to get into politics here, but since you brought it up -- don't you think that if it could be shown how the LN could operate in a decentralized (at least to some degree) manner, it would mitigate concerns of those who favor on chain scaling?
---
for general interest:
http://www.opte.org/

Carlton Banks
Legendary
*
Offline Offline

Activity: 3430
Merit: 3074



View Profile
April 08, 2017, 05:31:37 PM
 #8

You've responded to me today (in this thread), and the vast majority of the year previous.

It's not my fault your example Lightning channel chain in the OP was not sufficiently funded and interconnected to provide enough fund liquidity to service 1 specific hop, you designed that, presumably without understanding the implications properly. That you got it wrong is an objective fact, not an insult.

You can tell when jonald's lying: his mouth moves (another objective fact, and hence not an insult)

Vires in numeris
jonald_fyookball (OP)
Legendary
*
Offline Offline

Activity: 1302
Merit: 1004


Core dev leaves me neg feedback #abuse #political


View Profile
April 08, 2017, 06:49:05 PM
 #9

still more interested in arguing/insulting than in having an adult conversation I see.  ok, back on ignore you go.

Quickseller
Copper Member
Legendary
*
Offline Offline

Activity: 2926
Merit: 2347


View Profile
April 08, 2017, 09:03:43 PM
 #10

2) What happens to A's money that
she sent to G if she never closes
her channel with B?
They remain in A<->B channel. In LN funds belonging to a channel never leave that channel.

Example of Initial state of your piece of LN, channels are enclosed in parentheses:
A(1:0)B(3:2)C(4:1)D(1:2)E(5:6)F(2:0)G

After A pays 1 BTC to F (intermediaries don't charge fees):
A(0:1)B(2:3)C(3:2)D(0:3)E(4:7)F(1:1)G
Each channel contains the same amount of BTC as before. Note, that A can't pay F again through this route, because A's balance is 0, and D's balance (in A->G direction) is also 0.
If G closes his channel with F, they will receive onchain their shares of their channel and that's it.
I don't think that LN node operators will let any of their channels get to a state in which the counterparty's balance is 0 because this would result in a situation in which the counterparty could attempt to close the channel in an "old" state, and would be risking nothing in attempting in doing so.
stdset
Hero Member
*****
Offline Offline

Activity: 572
Merit: 506



View Profile
April 08, 2017, 09:13:44 PM
 #11

2) What happens to A's money that
she sent to G if she never closes
her channel with B?
They remain in A<->B channel. In LN funds belonging to a channel never leave that channel.

Example of Initial state of your piece of LN, channels are enclosed in parentheses:
A(1:0)B(3:2)C(4:1)D(1:2)E(5:6)F(2:0)G

After A pays 1 BTC to F (intermediaries don't charge fees):
A(0:1)B(2:3)C(3:2)D(0:3)E(4:7)F(1:1)G
Each channel contains the same amount of BTC as before. Note, that A can't pay F again through this route, because A's balance is 0, and D's balance (in A->G direction) is also 0.
If G closes his channel with F, they will receive onchain their shares of their channel and that's it.
I don't think that LN node operators will let any of their channels get to a state in which the counterparty's balance is 0 because this would result in a situation in which the counterparty could attempt to close the channel in an "old" state, and would be risking nothing in attempting in doing so.
It's a simplified example, made up to help people understand LN. In real LN there will be fees, unspendable balance residues. But that doesn't influence whole picture much. If say you fund a channel with 1 BTC and 0.001 BTC is an unspendable balance residue, that doesn't worry us much.

Pages: [1]
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.19 | SMF © 2006-2009, Simple Machines Valid XHTML 1.0! Valid CSS!