Criticisms of the Lightning Network

[Rath_]

Regarding criticisms, the need to keep private keys online and regularly monitor channels limits the appeal.

Mobile wallets are a small exception here. Once you enable payment receiving in Eclar Mobile, the timelock for new channels changes and you have to appear online once every two weeks. It's not bad when you compare it to 144 blocks (~1 day) default value.

[Welsh]

It's not a matter of being noob friendly or not. There was no option to set fees for closing channels. There was only one option to close friendly or unfriendly (something like that, dispute or not dispute, don't remember). I choose the one with lower fees , but even so fees were not customizable.

O believe every wallet should let users customize their fees, on sat/byte.
Because this would really lead to lower fees in general, everyone would look for lower fees unless absolutely necessary.
Free market would really work better.

For now most users just use "priority" or "standard" or whatever navme the wallet choose, but that really doesn't let you control the fees.

They've made the mistake of looking to streamline, and simplify the service, but without giving the optional option for those that understand fees a little more. I don't think they should scrap the "priority", and "standard" options. I think instead they should keep these as the default values, but allow any user which wishes to customize their fee the ability to do that within the same window.

[LoyceV]

I believe every wallet, lightning or not, should let you decide which fee to use.i would use 1 sat byte to close the channel, and waiting 5 days would be ok to me.

I'm no expert on LN, but isn't there a risk of broadcasting an older channel state (by the other party) if the closing transaction doesn't confirm on time? From what I understand, you can't just wait for a very long time if Bitcoin on-chain fees go up.

[Westingcote]

The lightning network is a great concept in general and has better privacy. However I'm a big fan of transparency and I think the way the current Blockchain displays all transactions that have occurred is a good way of doing it. In the lightning network you cannot see what transactions are going on because they are not stored within the Blockchain. One of the trades off of this is people who know your address will know how much you are spending which could be a privacy concern to those that are trading often and dealing with strangers a lot. Also we should look at the way fees work in the lightning network. They are generally much lower than on chain Blockchain but if you are routing through others payment channels then they will probably request a cut of the transaction which some might not like. It works in the same way the current Blockchain works but instead miners are receiving the cut.

I'm going to try and go through most of the pros and cons of the lightning network and see who agrees with me. I'm not an expert at the lightening chain but am definitely watching the space as I consider it to be one of the most interesting off chain developments being worked on and because we only have the Blockchain to compare to I'll compare each point to that.

Blockchain Vs Lightening Network

Privacy:

Without a doubt the lightening network is far superior to the Blockchain when it comes to privacy although despite some of the criticisms with using one address it has no major trade offs for the improved privacy. Basically when you send a transaction via the lightening network there is little worry to it being recorded permanently on the Blockchain and if it were to get recorded it does not display sources or locations of the funds. One of my favorite things about the Blockchain is it uses a sort of onion like encryption method to nodes so that nodes cant monitor addresses and transactions which is a much improved version over any other implementation of nodes interacting with the blockchain that I have seen.

Confirmation Speed:

This is a big one and one of the biggest criticisms that Bitcoin has is that transactions take too long to confirm. However this is intentional to allow miners to use their hashing power to get cuts of the transaction and keep the network alive and circulating. The only problem with this is depending on the current average network fees you could be waiting days for your transaction to confirm. In comes the lightening network which offers instant transactions almost like your credit card in how quick transactions are sent. I say "almost" because we have to allow for latency on the network but this is generally zero and your transaction should be sent right away. The blockchain manages transactions differently and actually purposely makes transactions confirm before being sent. This is to check whether all nodes on the network are up to date and a double spend hasn't happened.

However there is a drawback to this. You could continuously keep on sending your transactions on the same payment channel and have they sent instantly and received instantly too. However if you actually want to deposit this amount onto the Blockchain remember that the lightening network is off chain and the Bitcoin won't be deposited or recorded to be on the Blockchain until you decide to deposit it by closing a channel. depending on network usage and a few other factors this usually takes around 10 minutes. This is where things get a little complicated too. If the user that you are sending too requires 6 confirmations because of their business practice or whatever you will still have to wait despite you streamlining the process by using a off chain solution.

There is also another problem when dealing with the lightening networks confirmation. Basically it requires two people to open and close a channel. Opening and closing a channel usually takes around about 10 minutes like previously stated however if one side of the transaction doesn't agree to close the channel then the channel will remain open and this could take up to two weeks to be resolved. Therefore anyone you deal with on the lightening network requires a little bit of trust when dealing with them that they won't just disappear and not close the channel because that could be a pain in the arse for both parties. I can't see why anyone would do this on purpose as it just slows the process down for both but I would think that this does happen.

Fees:
Bitcoin fees can become expensive especially when the network has been congested like in the past. This does not really happen on the lightening network as you are not paying miners to use their hashing rate exactly and instead you are paying for the use of payment channels which every payment channel that you go through will want a cut of the transaction. This is usually pretty low when compared to Bitcoin fees though.

Centralization:
Because of the nature of the Lightening Network it is at danger of becoming centralized. What I mean by centralized is there will be great benefit to those that send transactions on a daily basis to use the same node every time which in time people will see the need of creating what has become known as "super nodes".

The best example I have seen of this is if you were to go to a fast food restaurant and they have created a node for their store in the city. They would keep this node open so that customers could pay instantly without having to create a new channel every time. This could eventually evolve into a nation wide and potentially worldwide payment channel rather than smaller ones being made all the time. America might decide that all american citizens can pay their taxes through the "United States Payment Channel" which would mean greater centralization in nodes.  

I believe every wallet, lightning or not, should let you decide which fee to use.i would use 1 sat byte to close the channel, and waiting 5 days would be ok to me.

I'm no expert on LN, but isn't there a risk of broadcasting an older channel state (by the other party) if the closing transaction doesn't confirm on time? From what I understand, you can't just wait for a very long time if Bitcoin on-chain fees go up.

The last time I checked the Lightening network did not check for existing channels and it is possible to broadcast to a older payment channel. However the best practice would be to make sure after every transaction the payment channel has been closed unless you are sending transaction fairly regular to justify keeping that payment channel open. My recommendation is to always close payment channels regardless though to increase privacy and decrease the potential of centralization.

[ABCbits]

Fees:
Bitcoin fees can become expensive especially when the network has been congested like in the past. This does not really happen on the lightening network as you are not paying miners to use their hashing rate exactly and instead you are paying for the use of payment channels which every payment channel that you go through will want a cut of the transaction. This is usually pretty low when compared to Bitcoin fees though.

But you need 2 on-chain transaction to open and close channel, which means you details with fees again. If user only need to make 1-2 transaction in some time, then create 1-2 on-chain transaction makes more sense.

Centralization:
Because of the nature of the Lightening Network it is at danger of becoming centralized. What I mean by centralized is there will be great benefit to those that send transactions on a daily basis to use the same node every time which in time people will see the need of creating what has become known as "super nodes".

The best example I have seen of this is if you were to go to a fast food restaurant and they have created a node for their store in the city. They would keep this node open so that customers could pay instantly without having to create a new channel every time.

I don't see how it's centralized, you always could route your transaction through another node, even though i won't deny most likely these "super node" will have big advantage, unless they decide to charge high fees for routing.

IMO those "super node" doesn't mean centralization, but concentration and monopoly.

This could eventually evolve into a nation wide and potentially worldwide payment channel rather than smaller ones being made all the time. America might decide that all american citizens can pay their taxes through the "United States Payment Channel" which would mean greater centralization in nodes. 

Your example doesn't show centralization because "United States Payment Channel" is the destination & government don't force their citizens to route LN transaction through their channel, even though it could serve as big honeypot.

Besides, if government could force their citizens to route all transaction through their channel, they also could force them to use wallet which created & managed by government.

[Rath_]

There is also another problem when dealing with the lightening networks confirmation. Basically it requires two people to open and close a channel. Opening and closing a channel usually takes around about 10 minutes like previously stated however if one side of the transaction doesn't agree to close the channel then the channel will remain open and this could take up to two weeks to be resolved.

The minimum_depth (a number of confirmations) is usually greater than 1, so it takes more than 10 minutes to fund a channel. The timelock is negotiated before funding a channel and it is usually 144 blocks (~1 day). For example, Eclair Mobile clients become an exception if the user enables payment receiving over the LN. In such case, the timelock is extented to 720 blocks (~5 days) for a local uncooperative close and to 2016 blocks (~2 weeks) for a remote one.

They are generally much lower than on chain Blockchain but if you are routing through others payment channels then they will probably request a cut of the transaction which some might not like. It works in the same way the current Blockchain works but instead miners are receiving the cut.

A 1/10 th satoshi fee sounds better to me than a 1000 satoshi one. Miners already receive a block reward while node operators don't get anything.

[mocacinno]

I've seen this discussion going towards the financial side of running a lightning node... Just to make sure no urban legends get created, i decided to share my own statistics..

Just some background info: I've been running a lightning node from before reasonably good lightning node explorers were around... Basically from the moment the first relatively stable beta version of c-lightning was released (once a week i still build c-lightning straight from the head branch on github).
I've used lightning-charged on top of c-lightning and a homebrewn payment system on top of lightning-charged for a long time. Recently i retired my own code and lightning-charged, and put btcpay on top of my c-lightning node. I've created hundreds of invoices and actually received +100 payments (so far).

Now, for the actual, up-to-date, statistics of my node:

Code:

lightning-cli listfunds | grep channel_sat
         "channel_sat" : 41646,
         "channel_sat" : 20200,
         "channel_sat" : 1276,
         "channel_sat" : 27012,
         "channel_sat" : 0,
         "channel_sat" : 0,
         "channel_sat" : 105550,
         "channel_sat" : 0,
         "channel_sat" : 396,
         "channel_sat" : 5285,
         "channel_sat" : 0,
         "channel_sat" : 0,
         "channel_sat" : 0,
         "channel_sat" : 0,
         "channel_sat" : 0,
         "channel_sat" : 0,
         "channel_sat" : 0,

So... currently 17 open channels, 7 with funds on my side of the channel. Dozens of other channels have been closed since i started running a node.

Code:

lightning-cli listfunds | grep channel_total_sat
         "channel_total_sat" : 100000,
         "channel_total_sat" : 5000000,
         "channel_total_sat" : 16777215,
         "channel_total_sat" : 100000,
         "channel_total_sat" : 500000,
         "channel_total_sat" : 500002,
         "channel_total_sat" : 5000000,
         "channel_total_sat" : 500000,
         "channel_total_sat" : 9000000,
         "channel_total_sat" : 9000000,
         "channel_total_sat" : 4499630,
         "channel_total_sat" : 1000000,
         "channel_total_sat" : 5446466,
         "channel_total_sat" : 409293,
         "channel_total_sat" : 100000,
         "channel_total_sat" : 85894,
         "channel_total_sat" : 58108,

All of them have funds on the other side of the channel

Code:

lightning-cli getinfo
{
   "id" : "03301e633b25d769377bf75ce6b6ed2ec570270bc06c8c02bf33c5bd2aa47da098",
   "alias" : "mocacinno",
   "color" : "03301e",
   "num_peers" : 25,
   "num_pending_channels" : 0,
   "num_active_channels" : 17,
   "num_inactive_channels" : 0,
   "address" : [
      {
         "type" : "ipv4",
         "address" : "193.70.78.148",
         "port" : 9735
      }
   ],
   "binding" : [
      {
         "type" : "ipv4",
         "address" : "193.70.78.148",
         "port" : 9735
      }
   ],
   "version" : "v0.7.1-172-gdbc0265-modded",
   "blockheight" : 588181,
   "network" : "bitcoin",
   "msatoshi_fees_collected" : 0,
   "fees_collected_msat" : "0msat"
}

yup... That's right... A grand total of 0 msatoshi in fees made during the months i've been running the node  

[Wind_FURY]

Another criticism is, how secure is Lightning? Has it experienced any attacks of some kind? That should worry the users, because we simply don't know.

I believe the reason why no one has yet shown any interest in attacking Lightning, is because the total capacity is not that big yet.

How much in total in USD was commited to the DAO?

[Rath_]

yup... That's right... A grand total of 0 msatoshi in fees made during the months i've been running the node  

That's interesting. Xian01 had been posting his Lightning Network node stats in the Lightning Network Discussion thread before it was closed and he actually managed to earn a few satoshis. If I recall correctly, he was using LND with its built-in autopilot. Here's one of his posts.

Has it experienced any attacks of some kind? That should worry the users, because we simply don't know.

The Lightning Network experienced a DDoS attack in March last year. 20% of all nodes were down at that time due to the attack.

[ene1980]

I'm just curious. For a change, let us look at the negatives. What are your criticisms concerning the Lightning Network? Do you think it's likely to succeed?

Personally, I think it's likely to succeed, though I think it will take a lot longer than people are expecting. I think it's going to be difficult for the developers to make LN easy enough to use for the end users, to the point that they couldn't even differentiate if they're using LN or just normal transactions.

If you want to look at the negative aspects of lightning network there are many, the most important thing is that you cannot make a LN payment offline and both the parties needs to be online for every transaction. For micro transactions it is the best option but for larger ones i would not trust LN. Since the transactions are made outside the blockchain you literally have to trust a centralized authority who started that network, in short LN is all about centralization  .

[Rath_]

Since the transactions are made outside the blockchain you literally have to trust a centralized authority who started that network, in short LN is all about centralization  .

Who are you referring to as the central authority who started the network? There's none. The Lightning Network Specifications have not been created by a single person or company. There are many different implementations (LND, Eclair, c-lightning and more) which are open-source. Anyone is able to audit the code and set up their own Lightning Network node. Payments don't go through a central server. You can open a channel directly to a merchant so that you don't have to pay any fees for multiple transactions (routing fees are negligible anyway).

[ene1980]

Who are you referring to as the central authority who started the network? There's none. The Lightning Network Specifications have not been created by a single person or company. There are many different implementations (LND, Eclair, c-lightning and more) which are open-source. Anyone is able to audit the code and set up their own Lightning Network node. Payments don't go through a central server. You can open a channel directly to a merchant so that you don't have to pay any fees for multiple transactions (routing fees are negligible anyway).

Hope you did not understand what i was referring to, i never debated about LN being a closed end program, it is an open source code and anyone can review it, how about the trust when everyone can open an LN network, how will you determine whether the channel is legit, the importance of bitcoin comes when you do not need a trust from a third party and that is why i got attracted in bitcoin back in the day, if you need trust i call it centralized, there is no two way about it mate .

[Rath_]

how will you determine whether the channel is legit

You can't create a fake channel because the Lightning Network is strictly tied to the blockchain. How would you determine if the coins you received were not double-spent? You would run your own node. I would do exactly the same on the Lightning Network in order to watch over all of my channels. Both of these nodes have to follow some standards and rules which in LN's case are written down in the specifications. You have to trust any software unless you code it yourself.

[Carlton Banks]

the importance of bitcoin comes when you do not need a trust from a third party and that is why i got attracted in bitcoin back in the day, if you need trust i call it centralized

that's wrong

please learn something (anything) about Lightning before you post again

[DaveF]

So I expect Lightning's perceived UX issues to ameliorate a little, but mostly through familiarity-driven self interest.

I have been bouncing a reply to this one line in my head since I saw your post a few days ago.
The word perceived is wrong. They are there. Everyone wants their pretty GUI menus and buttons on everything now.

Your TV gets a lower review on AV site because the smart TV features are not pretty enough.

If you want consumer level use you need to have that. It is not a were in alpha stage we will get to it. It's now become PART of alpha stage work. It all has to be done together. Because if LN is finished tomorrow but the UI / UX is not done for another 9 months then guess what LN is not done.

Just my view.

-Dave

[gentlemand]

If you want consumer level use you need to have that. It is not a were in alpha stage we will get to it. It's now become PART of alpha stage work. It all has to be done together. Because if LN is finished tomorrow but the UI / UX is not done for another 9 months then guess what LN is not done.

One of the things that may be hindering its perception is that most people who arrived in BTC turned up when the hard work was done. There's still plenty of tweaking but you can dive right in and it's all primed and waiting.

Lightning Networks are developing in real time in an organic way. Most people have no experience of that. The stuff they use every day is launched by a company when it's ready and it's immediately fully operative.

It's possible old hands will still be ignoring it thinking it isn't ready when newbies are batting their satoshis around without a second thought. Maybe someone should organise a cheesy Bitconnect style gala launch just to pander to human nature when consensus is that it's good enough.

[Carlton Banks]

Lightning is in a similar place development-wise that Bitcoin was in 2010 to 2012: usable product, and rough around the edges, but also moving at a decent pace.

I guess the difference was that the Bitcoin devs were "running out of road" and so needed to make sure that didn't happen before it got popular. Lightning devs have a whole network of highways, but the quality of the roads, maps and the direction signs is still a bit lacking.


@DaveF

if you want a GUI, there's:

• Lightning Labs desktop wallet (forgot the exact name :/ )
• Eclair mobile wallet (Android)
• That "Blue wallet" one (Android + maybe iOS)
• A GUI for c-lightning (pretty simple, looks like the standard Bitcoin GUI)

[arbiter5]

Another criticism is, how secure is Lightning? Has it experienced any attacks of some kind? That should worry the users, because we simply don't know.

I mean, Lightning is still software. Software bugs and problems are pretty much inevitable; even with Bitcoin itself. It just depends on how drastic the problem is, and or how quickly it gets resolved.

[btc_enigma]

Hi,

The biggest issue for me is  that you cannot have a  safe backup of your lightning funds using seed words

A safe offline backup solution is very important, if you want users to store any serious amount of funds on LN

[Carlton Banks]

The biggest issue for me is  that you cannot have a  safe backup of your lightning funds using seed words

no, the wallets used to manage Lightning channels do have a seed in every implementation


there's no seed for the channel addresses, but that's not relevant, as the funds will always end up back in the addresses of the underlying bitcoin wallet should a channel be closed.


think of it like this; Lightning channels are like a checking account, for regular comings and goings of cash. saying the channels don't have seeds is like complaining your debit card didn't come with a bank vault and an armed guard