Topic: Lightning Questions (Read 611 times)

Although i don't have a core knowledge of lightning network but what i understood till now is you can send through your friend channel as far it has a valid route and is connected with your channel.
Yes it allows channel relays and that's how the lightning network is working, because the payment can never be processed without the channel relays in indirectly connected network.

I work in real stores. Next to my stores there are several others. Each one uses its own payment structure with credit cards and cash registers. In this case, if bitcoin was to be accepted, would it be interesting for all stores to open a single payment channel? Would it be the most economical solution? Or in stores is it always more interesting to have a single channel?

Here is a well layed out article that explains how Lighting Network works in lamens terms. I highly recommended reading this! https://medium.com/@melik_87377/lightning-network-enables-unicast-transactions-in-bitcoin-lightning-is-bitcoins-tcp-ip-stack-8ec1d42c14f5

LND seems like the go-to implementation of LN

Is there any benchmark we can see to compare performance/functionality of the different implementations?
I want to eventually make a service instead of using LN as an end user, and I'd want to see what's more appropriate

Currently the network is still fairly small. But if you look closely, the network is really not a hub and spoke network. Even if many people have channels with these "main hubs", many of those nodes also have channels with other nodes too. The "hubs" thing right now is probably because there are only a few services that accept LN on mainnet and people are just connecting directly to them in order to pay.

I suggest you take a look at how LND does autopilot mode. It takes in various inputs and uses a model to figure out what channels are the bets to open. You can read about it in the release notes here: https://github.com/lightningnetwork/lnd/releases/tag/v0.3-alpha

LND does not need much more resources than those that are already required for running a full node (because you need a full node). You can also not run a full node and instead run it in SPV mode which uses BIP 157.

The current protocol limits channels to 2^24 (16777216) satoshis. There is no minimum funds; you can open channels or request channels to be opened where you provide no funds into the channel. IMO putting 0.01 BTC in a channel is reasonable.

At least on testnet, there doesn't seem to need much maintenance. The available software are capable of handling things automagically.

I wonder if there is a chance of c-lightning being merged into core codebase?

The software is still in alpha. GUI will be better.

Adoption is easy once a foolproof wallet comes out.

Swap to eth that got hacked (remember dao) numerous times before while btc had zero issues? No thanks. I trust the core team more than Vitalik Buttcoin.

Let everything aside, eth has a CEO while bitcoin has none. Satoshi left the game but vitalik is still around can manipulate the markets with his words. If this is not centralization I don't know what is. Even Charlie Lee realized his mistake and sold his stash lately.

It's a pipe dream. Currently, it doesn't even work, and it has SEVERE drawbacks:

1) payment channels still rely on the blockchain ultimately, so you will need higher throughput eventually anyway
2) it requires adoption - if it sees adoption similar to SegWit levels you're looking at nearly no throughput increase and high fees to open, refund and close channels
3) terrible UX

also: ETH HAS PAYMENT CHANNELS TOO

Swap your BTC to ETH - thank me later!

I consider joining in and be #reckless. I have some design questions though. Under the premise that I mostly dont want to spend funds and want to help process transactions and establish cheap path's along the network. Currently the topology[1] looks to me like there are a couple main hubs and everyone else just attaches to them.
#1 (how connect?) What would be a good approach to establish alternative paths? Connect to all bigger hubs, e.g. all with x+ connections? Or rather establish a connection between other nodes within the network that are not already connected via a single hop?
#2 (resource requirements?) What (additional) resources are required? Has anyone data on this? I consider upgrading my server anyway, so this would be a good switching point.
#3 (minimal funds/channel?) The current total capacity is only(?) ~3.68 BTC (again according to[1]), what would be - in your opinion - a good amount to put aside per channel? As much as possible obviously, but at which point does it become too little? Considering I plan to invest X amount of BTC this "minimal amount" would limit the number of channel I could reasonably keep open.
#4 (is much work?) How much maintainance is needed? Funds would move between channels, but the main flow may be in a certain direction along the path. Because people just buy stickers instead of sending funds back and forth. Has anyone any experience with this yet?

Thanks in advance.

[1] thanks for the awesome graph here https://lnmainnet.gaben.win/ btw

Yes. Opening and closing channels are both just on-chain Bitcoin transactions -- like sending coins, but with a different script obviously -- for which the regular transaction fees apply. Once a channel is open you only need to pay a negligible fee to each hop along the route of your payment.



Not sure about mainnet, but you can try the Eclair testnet android client or the HTLC.me testnet web client. Both are rather simplified in their functionality though, ie. HTLC.me hides opening / closing payment channels from the user and Eclair can only send, but not receive Lightning transactions. From what I've gathered so far you currently have to run a full node yourself to securely receive Lightning transactions.

Is there a way to run lightning without running a full bitcoin node?
I want to test a few things without having to use up a lot of bandwidth and space

This is more interesting to know how it all comes together.
Cause have heard of it being implemented but didn't know how it worked until reading this thread.
This initial opening of the channel and then someone filling the send to complete it's transaction is really going to solve all the hassle that has plagued bitcoin in the past 6-8 months since it's rise to stardom on television media outlets such as CNBC and Bloomberg in the recent months.

Awesome stuff. So the person opening the channel is the only one paying the fees?
If that's so that is going to be great.

I'd want to eventually figure it out and implement it on eCommerce solutions

Well, the way I understand it, it can only be opened and closed by the person that innitiated the Channel. They do not have any more control over it than that, and other people using these channels has even less access to do something with it.

I wonder how much control the developers have, when something goes wrong. At the moment miners can fast track specific transactions, because they are on-chain, but these transactions are off-chain.

Lot's of research still has to be done on my side, to fully grasp this.

No. Just No.

Read the lightning paper: https://lightning.network/lightning-network-paper.pdf

How does this lightning network work and by extension other lightning networks that will come in.

To keep it decentralised, will every node runner mine every transaction in their pool.
Is there some sort of order transactions will be stored by (ordered transaction hash numbers or something)?

68 nodes and 131 channels, just few more zero's after those numbers and we're ready to go. 

He paid 119 Satoshi/byte. Currently 73 Satoshi/byte is enough for a fast transaction, it's been relatively low all day. If it drops even more tomorrow (as most Sundays), I'll consolidate some of my inputs.

When you're making a transaction anyway, adding another output doesn't increase the size much. I'm not sure how many bytes it takes to fund LN, but his total transaction with 1 input and 2 outputs was 257 bytes. At current fees, I expect around 5000 Satoshi to be enough to fund a channel. That's less than $1 (on top of the fee you're paying anyway).

Very nice to see, thanks for the link. I'd love to experiment with it, even if it's just sending some money back and forth for testing.

---

Bitcointalk should have a Lightning Network board, it will be there eventually anyway, so why not start it already?

I agree on you too, it is really astonishing if it will be it's mechanics. I mean it's a too good to be true tech way back then but now it's implemented. But somehow i just think back if that's the case if it's more prone on open channel, does it mean it's also prone to future breaches or frauds? Just wondering..

No it is real.

Segwit really reduces the fees. A lot. Today i moved all my funds from my legacy address to my new bech32-segwit address and i just simulated a tx while writing this message. A 100sat/byte tx costs 1.25$ (0.00011btc). 200sat/b can get confirmed almost instantly which is $2.5 indeed.

Mempool is also being cleared today.

There are people buying real, physical things paying with lightning already. Take this guy for instance, he paid $2.50 to fill the channel:

https://www.reddit.com/r/Bitcoin/comments/7rkunw/lightning_the_future_just_arrived_at_my_doorstep/

The problem is when you run out of money, you must refill the channel with another on-chain transaction, so this incentives people to put as much money as possible at once on the channel, something that may not be viable for most people. For example if you want to buy a router, let's say it costs $150 bucks.. well most people isn't going to put more than that, so you would already spend all of your channel's funds at once not saving money.

I wonder he did get such a low transaction for the on-chain one tho, $2.50 is super low even if you use segwit. He must have gotten lucky with a period of low mempool activity.

LN is already up & working btw 

https://lnmainnet.gaben.win/
125 channels / 65 nodes on mainnet now.

Still, don't be #reckless wait for a stable idiot proof release.

Yes, if you already have a direct channel opened to your friend. However you don't get to choose which route your payment takes. It may cross your friend's node, but it may also take an entirely different route.



As mentioned by Kakmakr, the whole idea behind Lightning Network is to have a multitude of interconnected nodes (ie. wallets with open channels) form a network of payment channels that can relay transactions without requiring 2 parties to be directly connected to one another -- as long as there is a viable path that transactions can take ie. as long as they are at least indirectly connected. That's the Network in Lightning Network.

That is how it was supposed to work, is it not? The more channels that are opened, the easier and
cheaper it will be to find the shortest route for your transaction. As long as the channel is funded, it
will stay open for others to use.

I still have a very basic knowledge on the subject, like I think most people do, but I have a general grasp of the use case.

This is truely a mind blowing technology. ^smile^

I think we can discuss lightning questions here, so we can easily find them and address them in one place.

I have a few questions myself:
I heard Lightning allows sending across extendible networks, does that mean you can send through a friend's open channels?

In that case, could that allow for something like channel relays that could connect everywhere?