The Lightning Network

The Lightning Network, now coming online, is being touted as the savior of the increasingly sclerotic Bitcoin infrastructure. It is not a new coin, but rather a way of using Bitcoins, and purports to solve, or at least mitigate, several of the most grievous problems Bitcoin is exhibiting as it matures: rising transaction fees, agonizingly slow transaction confirmation, severely limited scalability, and exponential growth in the cost of maintaining the blockchain.

I’m only going to summarize enough here to make some points about the project. You can get a nice rundown in layman’s terms from Wired.  Probably the best source, if you have some technical background and not much time, is the Lightning Network’s page, which has a recorded presentation that summarizes the underlying mechanisms at a level somewhat above the cryptography. If you really want to understand the gruesome details, try the original paper. The cryptography is not developed from the ground up, but if you have the basics already, it’s very readable. There’s also a brief piece in Bitcoin News about the first real-world purchase using Lightning Network, which happened earlier today (January 21, 2018.)

What Is The Lightning Network?

The Lightning Network, which is not to be confused with Lightning Bitcoin (LBTC,) sits between Bitcoin users and the blockchain, buffering transactions in private pairwise channels that are set up between parties. After a pair of users sets up and funds a channel with Bitcoin, any number of subsequent transactions can take place between them almost instantly, skipping the delays and fees of blockchain interactions.

Fees

There is a blockchain interaction with an accompanying transaction fee incurred for setting up the channel because coins must be moved to the control of the Lightning Network’s protocol. Subsequent blockchain activity in behalf of the channel, however, is normally confined to just two occasions: when additional funds are added and at the end of the channel’s existence. The idea is that the blockchain fees will be amortized over many free or nearly free transactions between the two parties over the life of the channel, reducing the expense and sparing the blockchain much processing.

While the channel exists, the details of transactions going back and forth between the two parties are normally handled entirely within the channel in such a way that when it is all over, the only thing that must be recorded in the blockchain is the net result, i.e., the final ownership of funds that got put in.

The Final Authority

The blockchain remains the final authority and there are some situations in which participants may opt to bail, kicking a problem out to the blockchain, but these are mostly to resolve problems; the main point of the scheme is to reduce blockchain interactions.

What Is a Channel, Anyway?

A channel is not a server or other physical thing. It’s not even a computer program. It is the cryptographic embodiment of a relationship between two parties. As such, it is more like a contract, except that it dynamically interacts with the two parties, and can take certain other actions as well that we’ll touch on later.

Channels are created holding a quantity of Bitcoin supplied by one of the parties. Bitcoins aren’t really things, so what this means is that the channel, not the participants, is what has the right to post a final statement about who owns that value to the blockchain. It also has a lot of cryptographic odds and ends to ensure that both participants can interact (or fail to interact) in a secure way, meaning they can’t cheat, outsiders can’t steal, and all of the numerous ways that things can go wrong are covered, e.g., if one person dies or goes away on vacation, the coins don’t languish forever.

For as long as the relationship persists, the two parties must agree to any changes to the ownership status of the coin. (The humans involved don’t have to understand most of this, of course—it’s all embodied in the protocol.) If they agree to move some of the value from one to the other and both sign off, then the accounting in the channel reflects the status change. If they can’t agree, the status remains the same.  Channels are bidirectional—both participants can own value and move or not move any amount of it to the other.

Distribution

One extremely nifty thing about the scheme is that it requires no global central authority like Mt. Gox that must be trusted by the participants to hold the Bitcoin for them and arbitrate differences. Not only is there no trusted central service, there is no such service at all. It’s completely distributed and therefore, there is no intrinsic bottleneck to scaling other than the blockchain itself, and a big part of its purpose is specifically to minimize work for the blockchain.

People Aren’t Always Reasonable

The protocol can’t make people be reasonable, but if at some point the parties can’t agree on how to make the next transaction and become deadlocked, an unreasonable or unresponsive person can keep things stuck forever. Channels are created with a defined lifetime, and eventually, the channel will simply expire and ownership will be reflected in the blockchain in its most recent agreed upon state. For example, one party might lose his or her mind and decide that “it’s mine, all mine, he he oo hoo ha ha.”  The crazy person can’t hold the other’s coins hostage forever because if they never agree, eventually the channel will expire and the division of coins will appear on the blockchain as it was most recently agreed upon.

The Catch

The catch, as alluded to above, is that to make this work, users do have to commit coin to a channel in advance. It’s sort of like pre-paid gift cards; you commit funds to the card (i.e, a channel) and then you can spend up to that limit. That’s why the payments can’t bounce like checks.

Kevin Bacon

Any system where you must tie up money in advance to cover all of the transactions that you might do over an extended period, for every business relationship you have, would be a hard sell.  For this reason, the protocol allows for networks of channels. The committed coin can be shared among payment channels, and equally importantly, you don’t need to open a direct channel with someone whom you wish to pay, so long as they are connected transitively to someone with whom you do have a channel. In other words, you only have to have a channel with someone who has a channel with someone who has a channel, etc., etc. until the chain gets to whomever you wish to pay. The protocol will make sure that right person ends up with the coin being deducted and that every channel along the way agrees with what’s happening.

This networking makes the model more like a pre-paid credit card that works for all of the counterparties in your extended network, except that there is no underlying credit card company that can pull a fast one.  One caveat is that the counterparties that you wish to do business with may not all be transitively connected to you, so in practice, you may need to maintain and fund multiple channels in order to cover them all. Widespread adoption of the Lightning Network should mitigate this situation through the six-degrees-of-Kevin-Bacon effect.

Hold On Just A Minute

Lightning Network provides a nifty solution to a lot of Bitcoin’s problems, but let’s not forget the original problem Bitcoin itself was invented to solve: if you wanted to pay someone over the Internet, then they had to either trust you or you had to involve a trusted third party like PayPal.

Solving this problem is what makes Bitcoin digital cash—with Bitcoin, you and a counterparty can do an electronic exchange with all the certainty and anonymity of handing over a stack of Benjamins. Just as with Benjamins, there is no question of the payment not being honored, or not having enough on deposit to cover the transaction, or a third party turning over the details of your transactions to the government.  It’s cash.

Is Bitcoin Over Lightning  Still Cash?

I felt a little defensive about this question. I wasn’t sure that Bitcoin over Lightning Network still qualifies as cash. After giving it some thought, I have to say it’s not.  If you have to form a prior relationship and set up what amounts to a pre-paid credit card, it’s definitely not cash.

It’s a trick question. Being cash is just a means to an end for Bitcoin; the point isn’t to “be cash” but to allow payments without requiring trust. Being cash is just a means to an end, and the Lightning Network, though not a cash system, preserves this. Nowhere along the line does anyone have to trust anyone else, just like using bare Bitcoin.

The Lightning Network doesn’t claim to be a replacement for Bitcoin or a better Bitcoin. What it claims to be is a better and scalable way of using Bitcoin. I own cash as well as a credit card and use either, as appropriate. Should you find yourself needing to pay for your kilos of Fentanyl in Bitcoin using the raw blockchain, you remain free to do so.

Is it Useful?

The cash thing turns out to be a red herring. A better question is, is it is useful?

It’s only beginning to be tested, but my intuition is that it certainly is. The specific benefits depend upon what you are doing.

Fees

Transaction fees currently (i.e. prior to Lightning Network) make small payments using naked Bitcoin impractical, but they have never really mattered for big purchases; a fifteen buck fee is noise if you’re shifting around ownership of old masters, antiques, or gold bricks under the radar.  Fedexing paperwork costs more than that and takes overnight, minimum.

Time

The time it takes for big transactions to clear the blockchain does matter a great deal, however, for many reasons. For instance, it matters if the principals’ lawyers or agents have to sit around on the clock waiting for the transaction to be verified. An hour delay is normal now (January 2018) but it has been known to take as much as 16 hours in extreme cases. For another thing, your counterparty might find the delay irksome or risky.  For yet another thing, the value of a Bitcoin can vary by thousands within the span of 16 hours, or even within one hour. For the typical large transactions, speed rather than cost is probably by far the biggest benefit.

Obscurity

There are people who value Bitcoin because the nature of their business requires extreme discretion. The last thing you’d think such folks would want is a system that involves setting up persistent payment channels with their counterparties. You might think so, but pairwise channels have an opacity that the bare blockchain does not. To my eye, Lightning Network looks like a boon to the paranoid.

This isn’t something that Lightning Network people or other experts claim, just uninformed speculation, but it doesn’t seem far-fetched.

Every transaction using a straight Bitcoin payment leaves in the blockchain a permanent, public, indelible, timestamped record of how much Bitcoin moved between which two wallets. A lot of people don’t realize that the transactions themselves are completely public; what makes Bitcoin anonymous is that the blockchain doesn’t record who owns the wallets that the coins belong to. The investigator who could somehow map wallet ID’s to specific people could discover a great deal.

Wallet ownership or information from which it can be inferred is exactly the kind of obscure tidbit that can figure in a deal with federal prosecutors years after the fact. It is also the kind of thing that big-data techniques let an adversary infer from miscellaneous low-quality data that either falls out of an investigation or is just out there in the ether. For instance, if you know the approximate time of transactions for known amounts you can match them to the blockchain and see which wallets fit. You don’t even need the amounts if you know a lot of times—just look for pairs that correlate with those times. Data doesn’t have to be complete or exact to make good guesses and then sharpen them with subsequent information.

The Lightning Network adds multiple dense layers of obscurity on top of Bitcoin as an almost unavoidable byproduct of how it works. First, only the net result of arbitrarily many transactions made on a given channel ever makes it to the blockchain, and the transactions can spread out over a long duration. This compounds the difficulty of extracting a meaningful story from the blockchain even if you know who owns what wallets and it makes wallet ownership harder to infer because the blockchain transactions don’t necessarily correlate with known times or amounts.

Secondly, the whole strength of Lightning Network is that it does not bottleneck on a single global ledger like the blockchain; the channels are lightweight temporary entities that come and go from individual computers; they aren’t centrally hosted by an entity keeping records. Channels aren’t ledgers; each transaction logically overwrites the previous state of the channel and when the channel closes out, it’s all logically gone. There might theoretically be information in logs of some kind, but even if there is, the whole thing is highly fugitive as even the nodes themselves don’t need to be logically permanent. It sounds like an investigatory nightmare.

There is no telling how this aspect might play out with regulators. It could be the dawn of a golden age of moving money opaquely, which would greatly enhance Bitcoin’s popularity with non-speculators, or it could be such a gruesome enforcement problem that the government could end up putting the kibosh on the whole thing. Or it could make no difference.

I’ve always suspected that the authorities might indulge Bitcoin precisely because it keeps a precise record of things people assume are secret. Information about who owns what leaks out in a thousand ways, and in this age of big data, who knows what a determined NSA or CIA might be able to infer over time by correlating many data sources?  To my untutored eye, Lightning Network looks pretty thorny for the authorities, worse even than Bitcoin itself.

Enabling Micro-transactions

The selling point you hear about most is micro-transactions. This one doesn’t excite me too much, but it’s the one that the Lightning Network people mention most.  “Starbucks” is the cliche example, but what else does a typical person repetitively spend petty cash on frequently enough to justify setting up a channel? The grocery store, rent or mortgage, the gas station, car-payments; I’m racking my brain to think of anything else I spend money on that isn’t a one-off and none of my regular payees take Bitcoin. Lightning apparently does a fantastic job with micro-transactions. The question is, is it a solution without a problem?

The Real Question

A more important question is, will the usefulness of the Lightning Network justify having to keep your Bitcoin tied up in channels?

Interesting

The last item above points at the crux of what’s confusing, not about Lightning, but about Bitcoin itself or any similar digital currency (this does not apply to Ripple, for instance, which is not a currency in the same sense.)

If a general purpose payment system made you lock up real money that way, it would probably be a show stopper, but Bitcoin isn’t money. “Huh?!” I hear you say, ” I thought you said it was cash?” Yes, it’s cash, but it’s not really money. Read on.

The confusion for a lot of people is that currency and money are not the same things. Currency—cash—is a special case of money and actually amounts to only a small fraction of the world’s total money. The majority of money exists in more vaporous forms such as debt, floats, deposits in banks, money banks don’t have, but may lend anyway, etc. Bitcoin is currency but not full-blown money because, unlike dollar-denominated currency, you can’t deposit it at interest in a money market fund, a CD, or put it in the stock-market, so that it can earn more money.

People get confused by this because they see profits getting made by holding Bitcoin. Yes, but those profits are changes in the dollar value of coins, not coins begetting more coins. It’s not that way with dollars. Like Bitcoins, they can just sit there under your mattress,  but if you put them in the stock market or the bank, your dollars can make more dollars.  Bitcoins can’t do that. Thus, dollars are money that may or may not be in the form of currency at any given moment. Bitcoin is always currency.

Bottom line: Does Lightning Make Sense?

From the perspective of the individual holder, the point is that Bitcoins can’t be invested in the stock market or put in the bank for interest, so if you “tie them up” in a Lightning Network channel you aren’t actually forgoing profits that you might have made otherwise, so long as they remain similarly accessible. The purchasing power of the Bitcoins in your channels is the same as the Bitcoins on your USB drive, and any value change is reflected in the prices you pay when you spend them, not a change in how many you have.

By owning Bitcoins at all, you’ve already accepted the risk that their value will fall while you own them; there is no reason to think that Lightning adds any additional risk, and it may even reduce one’s risk because it means the holder can execute transactions to convert coins to dollars, rupees, or old masters faster in the event of looming Bitcoin market-price troubles.

So the catch I mentioned at the beginning might not be a catch at all, especially if you can use Lightning to reduce the overhead of gambling on the exchanges, which is mostly what Bitcoin is used for at the moment. It’s not clear yet, at least to me, whether this will be the case, but it seems like it should be because dollars are no different from any other commodity as far as Bitcoin transactions are concerned. But take this with a grain of salt because there may be some impediment to their use on exchanges that I have not thought of.

From the perspective of an individual, it seems to be all good.

From the point of view of Bitcoin as a whole, Lightning Network makes a ton of sense. The overhead of confirming transactions is now all but unsupportable and is an increasingly a severe impediment to wider adoption. Lightning seems to mitigate this and increase usability.  For Bitcoin as a whole, it’s probably either onward and upward or bust, so adoption of the Lightning Network or something like it is a matter of do-or-die. A new lease on life.

The final perspective, that of the numberless hoard of new investors coming in, is the problematic one.

In my opinion, Bitcoin, however fascinating it may be as technology and cryptography, looks like a bubble, and it will go hard on anyone still holding them when the air goes out.

If that seems too grim, reflect that Bitcoin as a whole is a zero-sum game.  Beyond the hard core of cryptographers, computer scientists, and people with an ideological interest, the mass appeal is about getting rich. The trouble is simple math. For today’s investors to get rich, tomorrows investors must be proportionately more numerous. That dynamic requires an ongoing geometric increase in the number of new investors. Geometric increases generally can’t go on for long.

For this reason, a pessimist would say that the success of the Lightning Network, an extraordinarily powerful idea, in some ways more elegant than Bitcoin itself, would be a very bad thing if its success merely delays the bubble-popping that today seems inevitable. Today’s holders would have more time to get out winners, but at the expense of many more losers later on.

But that may be too grim. Bubbles must end, but the optimist would claim that they don’t necessarily have to pop. They can also deflate.

The Lightning Network could conceivably turn out to be a miraculous loophole if it can get Bitcoin out of the realm of speculation and let it finally become a bonafide currency for quotidian use. In that scenario, the owners of coins would no longer be “investors” but people who actually want to spend them. The coins held for speculation today would benignly migrate into actual use as currency and tens of thousands of people could keep their retirement funds.

I don’t think anyone can make the final call yet. It will be an interesting year coming up.