Lightning Network Flaws

https://github.com/davidshares/Lightning-Network

by skybrian • 5 years ago

75 47 5 years ago

skybrian2• 5 years ago

This is an impressively long list of articles about problems with Bitcoin’s Lightning Network. It’s not my area of expertise so I thought I’d see what people have to say about it.

__Admin__• 5 years ago

LN is a "soltion" Blockstream (tm) created to solve a problem Blockstream created.

"Solution" because it does not actaully allow fast and cheap "cash-like" transactions as Bitcoin was originally designed to do (and still can easily scale to Visa-like levels).

From day one Satoshi had built Bitcoin to scale to to Visa levels of transactions by simply incrementaly increasing the block size over time. Blockstream hijacked the Bitcoin Core GitHub repo and soon after disabled several core features and have since refused to increase the block size while pushing their products (Liquid and LN) as "solutions" to this problem they created.

Vecr• 5 years ago

No, Blockstream's thing was Liquid, an actual product that was supposed to make them money. Lightning can't really make them any money (they would get caught in a race to the bottom), unless they want to do consulting. Consulting probably would not give them much or any control.

3np• 5 years ago

The list is a mixed bag of past CVEs; discussions on challenges or concerns (some absolutely valid) with architecture, routing, emergent centralization and privacy; FUD social media/blogspam; past and current (but solvable) UX issues; reddit/twitter rants; etc. Many of the links are nothing-burgers but there are a few exceptions.

Specifically wrt to privacy, there are pitfalls and subtleties in using Lightning Network privately but it is doable. I think this article does a good job summarizing the current state [0].

IMO LN is secure enough to sleep well at night with open channels but you definitely want uptime monitoring and at least one watchtower hosted on separate infrastructure from the LN node, or great trust in whatever custodial wallet provider you may be using to not doxx you if you do and that's a concern. (Seriously though, any cryptocurrency going through any form of custodianship with any form of company/business will eventually have to be KYC’d and accounted for to authorities, if currently existing and planned regulations take effect. Self-custody and, for L2 like LN, self-hosting, takes some self-education but is the only way to get the true benefits of cryptocurrencies)

[0] https://abytesjourney.com/lightning-privacy/

mw888• 5 years ago

Weird to use Github for a simple list of hyperlinks. Anyways, this debate is extremely political as this is essentially a Bitcoin Cash viewpoint, and lightning makes BCH obsolete if it works, so from any dogmatic holder of BCH the motivated reasoning is obvious. That doesn’t mean issues can’t be true, but from following both the BTC and BCH communities for a long time I have trouble stomaching the same critiques about lightning over and over when it’s clear one community has never wanted it to work, and actively shit on progress or innovation. As soon as you see the monetary incentives posts like this start feeling like an insult to your intelligence, whether they fundamentally have a point or not.

skybrian2• 5 years ago

Good to know that it’s a political minefield. It doesn’t help to understand whether it works, though?

Maybe someone who has actually investigated will have something to say?

lawn3• 5 years ago

At the same time, if LN doesn't work then BTC is completely and utterly obsolete.

And finally, it's not just "a Bitcoin Cash viewpoint" and labeling it as such does the criticism a big disservice.

idiotsecant• 5 years ago

Not to say that you're making this point, but I would argue that if Bitcoin is obsolete and irrelevant than Bitcoin Cash most certainly is equally irrelevant and obsolete. The Bitcoin Cash approach to distributed database design is kind of like seeing a car with a 1kw engine, realizing that it will never be able to reach orbit, and putting a 100kw, diesel-electric hybrid engine in it instead. I think it's undeniably better, but in a direction orthogonal to what the protocol needs for long term usefulness - a more sustainable consensus mechanism with massively higher throughput.

eluusive• 5 years ago

"It's political" is true; but it's also true that it doesn't work as advertised.

aww_dang2• 5 years ago

Attempting to debunk critiques because they have political implications is political in itself.

If the opposing view is political, then the proposition is political. Dismissing it out of hand because it is 'political' becomes nonsensical from this point.

LN has been presented as a panacea to BTC usability issues. Are there no incentives in this promotion? Why would they be beyond critique?

Yes, technologies deserve a chance to develop and stumbling blocks along the way can be expected. However LN has been a basket case of never ending issues. There are other platforms aside from BCH which are effective for transacting. I won't list my favorites in the interest of avoiding controversy.

mattwilsonn8883• 5 years ago

You would make a good point, except that this post is not anything approaching an argument - instead its a list of links of other arguments that the reader has to independently go through and judge. The quantity of opinions says nothing of the quality of them, and the political angle explains why such a strange format was used.

dkobia• 5 years ago

Considering the author seems to be committed to BCH gives this list a somewhat jaundiced perspective. The LN vs BCH arguments are quite fervent mostly because it is likely a zero sum game - the bitcoin purists vs the pragmatists.

mattwilsonn8883• 5 years ago

I feel I have to defend Bitcoin a bit here - its not like the Bitcoin fork most popular today, "Bitcoin," is the result of just purism - there are legimitate reasons for stick with small blocks, namely, optimizing for decentralization which is pretty important for the pilot run of crypto in a world where corps would love to have control over these protocols.

Of course BCH folks would say that corps have hypnotized the BTC community and convinced them that small blocks are good, when in reality they just want to cripple it so it can't compete with traditional payment methods. The main issue I take with BCH folks is they act like smaller blocks are worthless - if they at least addressed the merits with some earnestness it would be easier to see them as rational in their reasoning.

freeAgent• 5 years ago

“If it works” is kind of important to determine, though. From what I’ve seen, LN frequently doesn’t work and transactions fail for unknown reasons (offline nodes, no route, liquidity issues, inability to open new channels with peers, etc.). The “solution” to LN’s problems right now appears to be custodial, centralized wallet providers, but that defeats the purpose of using Bitcoin. For example, Twitter is using Strike for tipping, but that wallet is custodial and requires KYC to use. That’s not aligned with what Bitcoin is supposed to be, IMO.

lottin• 5 years ago

This is because "peer-to-peer digital cash" is incompatible with features that consumers definitely want, such as, reversibility, legal protection and customer support.

ilaksh• 5 years ago

Take a look at Algorand if you are looking for a high-performance cryptocurrency.

rvz• 5 years ago

Very well then. I will read its whitepaper. Nice to see other cryptocurrencies emerge but some of them will fail.

baby• 5 years ago

I made a video to explain parts of their consensus algorithm https://www.cryptologie.net/article/484/algorands-cryptograp...

Konstruct• 5 years ago

The biases and cultish mentality of hardliners in the Bitcoin community make it difficult to have an honest discussion about these issues. There are interesting and nuanced security flaws listed here around both DoS and more recent routing fee exploits. My takeaway here is that there is a general lack of understanding and sufficient security testing around LN implementations. I wouldn't run my country on it but I wouldn't throw it all in the trash and abandon the concept like the author of the Github link is suggesting.

mattwilsonn8883• 5 years ago

That's a very manipulative comment to make. I don't own either, I am impartial, but both sides have their own money invested in what they see as the right way forward for a new technology. Both sides are at times horribly biased and I've seen more reasonable people on the BTC side by virtue of it being a larger and more well funded community.

ricardolopes• 5 years ago

Like others mentioned, this list is a mix of some valid known issues and some low quality bad-faith bashing filled with incorrections. It's also weird that it goes all the way back to 5 years ago, given how young this technology is and all the improvements it has seen since then. Anyway, I'm glad Lightning has been gaining momentum lately. Lots of innovation happening, and lots more still to do to address its current flaws.

user-the-name• 5 years ago

lots more still to do to address its current flaws

That is a bit of an understatement, given that actually implementing a routing algorithm that works at scale is an open problem in computer science with nobody even knowing where to begin.

gwicks56• 5 years ago

doesn't the internet route at scale?

orena2• 5 years ago

There is a solution to that, if they remove the making on channels in/out funds the problem is solved...

varajelle• 5 years ago

So this is a page made by a lightning network "hater", who just collected every bad article about it out there.

joshsyn2• 5 years ago

So you dismiss by attacking his character instead of refuting his claims by evidence?

xiphias22• 5 years ago

,,WARNING: If you try to use the Lightning Network (LN) you are at HIGH RISK of losing funds and it is not recommended or safe to do at this time or for the foreseeable future.''

This looks outdated. This warning was true a few years ago, but nowdays the risk is low enough for everyday use (as a checking account, not as a savings account).

jeroenhd• 5 years ago

How is any system that risks spontaneously losing funds any better than just using the existing banking infrastructure that has this down to an absolute minimum?

Of course market speculation has made any cryptocurrency system completely unusable for long term value storage so I wouldn't put my savings in there anyway, but even for a checking account the existence of risk of spontaneously losing money that you can never get back while using the system normally seems like a massive design flaw.

xiphias22• 5 years ago

It's never spontaneous. Banks and transaction processors lose money regularly, they just don't expose merchant fees and chargbacks to the buyers directly, but when you remit money to another banking system or want to cash out your balance, you are exposed to those fees and times and limits.

Bitcoin and lightning network exposes the end users to all the potential security issues, but gives finality and no bank bailouts in return.

ogogmad• 5 years ago

Are the security problems with Lightning fixable without resorting to centralised custodians? If people need Paypal to transact with Bitcoin, it undermines Bitcoin's selling point of decentralisation. Might as well use Paypal to transact in gold, which has the same economic properties as Bitcoin, but doesn't involve expending enormous amounts of energy doing Proof Of Work.

I've never owned BCH, and I was once an owner of BTC who steered clear of BCH, but I have to admit that the BCH solution has the major advantages of simplicity and tried-and-provenness. It's tried-and-proven because it doesn't make any additions to the algorithm Bitcoin uses, which after 11 years of experience we know works. If you want a computationally secure system, then you want it to have exactly the aforementioned qualities. I guess if gigabyte/USD prices for storage devices were to drop fast enough, and if it were possible to turn down the latency of transactions without compromising the safety of the network, then maybe the Big Block solution could work? I'd still be worried about surges in latency when there are increases in demand.

The LN approach seems very complicated and experimental, and involves playing with real money. It's been in development since 2017 and is perpetually 12 months away from being ready. Some papers introducing attacks list numerous difficult tradeoffs that need to be made to make the network secure against them. The descriptions of it I've seen online seem naive and simplistic compared to the reality. Complexity is antithetical to security and correctness.

dmm• 5 years ago

because it doesn't make any additions to the algorithm Bitcoin uses

I haven't paid attention to BCH in a while but when it was released they had changes to the difficulty calculation that allowed difficulty to drop before the normal 2016 block interval.

In principle this could give miners greater control over inflation.

I'm not saying it's good or bad but it is a change to the bitcoin algorithm.

smoyer• 5 years ago

Was the algorithm changed or just a constant/variable. I'm not as familiar with how mining difficulty changes but the reward halvening is based on a relatively simple algorithm that uses a constant of 210000 (blocks). Changing that constant doesn't change the algorithm but certainly changes the behavior of the system. In this case, the security of the system is clearly not changed when the constant is changed.

My understanding of mining difficulty is that you could tune it without compromising security so long as it's hard enough to provide dampening to the system. Right now it's adjusted to produce one block every 10 minutes on average.

Ultimately, if you want to keep the algorithm you need to either allow bigger blocks or allow blocks to be produced more often.

lawn3• 5 years ago

This is a necessity for a minority chain, as it has to adjust the difficulty more often than Bitcoin, otherwise large variance would cause the difficulty to rise so much that it wouldn't be profitable to mine, effectively killing the chain.

eraad• 5 years ago

This is just a list of links with no context or real technical analysis. I guess you are comparing LN to something else, in that case you should too make a list of links of the problems and fraud issues with, for example, fiat based payment networks or as another example, real adoption issues for things such as bcash.

And if your objective is to discard LN because of those issues, those same issues do not seem to matter in the real world where people are actually using LN for real transactions every day all over the world. This is public data you can check yourself.

Anyways, I think this list of links is biased and doesn’t add value to someone trying to understand modern payment networks.

DennisP2• 5 years ago

Seems to me the most critical flaw is that you have to be online to receive funds, unless the sender is directly connected to you instead of hopping through intermediaries. This is inherent to the way the whole idea works[1].

Your only options are to either keep a lightning node online yourself, or to trust another node to custody your funds. At scale I think most people would do the latter, which sorta breaks the whole point of "p2p digital cash."

[1] https://github.com/moneybutton/yours-channels/blob/master/do...

orena2• 5 years ago

Not true - https://github.com/breez/LightningRod

schemescape• 5 years ago

Can you clarify how that is different from what the parent poster said?

or to trust another node to custody your funds

sjtindell• 5 years ago

Who can look at defi exchanges, bridges, and wrapped tokens, and think that any particular coins technology is going to matter?

Lightning Network Flaws

Related Stories