Ask HN: What problems does blockchain solve? People seem to be raving about blockchain technologies and new startups are popping up to apply it to new industries every week. How does blockchain actually help a sector like healthcare, finance, retail, etc? Are there examples of where having a decentralized data store solved major pain points? |
To know why that matters you have to understand the problem(s) it's solving - it's something people have been working on since the beginning of money, but there are so many edge cases we usually ended up just going back to a centralized authority. In the past you've needed something like a bank or a government or a company to have the authority to declare, with authority, "Person x owns the title to that house."
There are a few problems with that: They could do things to screw everything up (a government printing currency and causing inflation), be corrupt (countless examples), or perhaps not even exist because it's too much effort to create, set up, and monitor some body.
With the blockchain, the software everyone runs keeps track for everyone else in an incorruptible way (yes this is a simplification), so you're free to make changes (or transact) without having someone in the middle to wait for, and with no one that can really screw things up. Just boom, instant marketplace.
The most obvious thing you'd trade on a blockchain is some kind of asset - a cryptocurrency is the main example, or titles or stocks or something valuable.
Say, for example, I have some loans that people owe me on. I can throw those loans on the blockchain, and without a clearing house somebody else can buy them, and when the loan payment comes due you can easily see who owns the right to be paid.
It seems trivial, but allowing people to seamlessly make transactions of any kind without the overhead of a bank or clearing house is a big deal, and has a ton of applications.
If we use Bitcoin as an example, it magically just works by consuming a pretty extravagant amount of electricity for the miners/nodes to run.
Of course bitcoin doesn't do everything those banks do yet, but on a per-transaction basis it uses orders of magnitude less resources than the current system
Are any of those adjectives true? Its no where near instant, it can be "corrupted" by an entity with enough hashing power and there isn't any magic to its operation.
Cutting through all the [templated] legalese, we can say: "Contract A, between parties K and L, concerning transaction T: T shall be considered done and legally binding once Coloured Coin X has been paid from address [foo] to address [bar], and has been verified by no less than 20 subsequent blocks."
The idea of a Coloured Coins is that you use something uniquely identifiable to make a transaction in globally verifiable transaction log. (Iow: blockchain.) Real-world contracts can reference the binary result of "has this transaction happened?" to verify that a contract has been completed and an off-blockchain transaction is valid.
How to trust a single entity by using a magnitude of peers to establish an agreement. In the case of Bitcoin it's a monetary transaction.
One example would be file digesting. When you consider malware: if files were signed and digested on ablockchain we would have better oversight of the authenticity of a download. A file could be confirmed against the blockchain to detect if it matches the certification. Since the source alone cannot be trusted, the confirmation of many adds that layer of trust.
Anything beyond the issue of trust is not suitable for a blockchain as there are more efficient technologies.
I have a book coming out next week (!) on the subject. https://davidgerard.co.uk/blockchain/ I have a whole chapter on business blockchains, and I looked hard for a real-world example of one in use. There actually aren't any.
The closest we have is .. git! Transaction ledgers, each with a tamper-evident hash, in trees and chains of hashes. Devs routinely throw entire ledgers/repos around, identified by hash. It definitely counts as a "distributed ledger technology", and it's hugely successful. The only thing it doesn't have is a consensus mechanism - it's "here is my tree" or "here is this repo, this is the hash".
As we're seeing here today, most claims for business blockchain are literally the airiest hypotheticals regarding Bitcoin, with the buzzword changed.
The usual concrete posited use cases are interoperability (that it will magically clean up your data and formats) and that it will magically resolve real-world human-level disputation. Neither of these is likely to work out that way.
"The decision of whether to use the technology as an commercial-grade replacement [of ASX's existing settlement system] will be made in fiscal year 2018."
The thing is, though, that blockchains promise replacement by magic, and there's no magic there - it's just as much work as any major system replacement. Also, other stakeholders are already unhappy this doesn't interface a lot like the old system did.
If this pilot works out at all, I suspect the first thing they'll do is rip out the blockchain and replace it with a conventional centralised database. Because this is for dealing with the ASX, and centralisation is obviously the correct thing and "permissioned blockchain" is another word for "the world's most inefficient centrally-administered clustered database".
The only use case I know, beyond proof of ownership of coins/tokens (and, again, how can that be applied to a real world problem), is proof of document state at a certain point in time (via storing document hashes in the blockchain) which clearly has some uses in law for example, but apart from that I am stumped.
️=> Most used application is money-based transactions. Banks are central authority and are affiliated with governments, which makes them charge of your money. Use blockchain (Bitcoin).
=> Supply chains are inefficient due to checks and sign offs by so many people. Use blockchain (Hyperledger)
️=> People don't agree on terms, and they end up in a feud or a lawsuit which is costly. Use blockchain (Ethereum)
Blockchain, and decentralization on whole is seen to be as future. Many want to tap it before others. Of course, large number of startups will fail, but it doesn't say anything about the potential this technology holds.
Because a shipping form gives you global ownership of a container and because corruption, the security required to handle shipping forms pre block chains meant it was more expensive to ship the form than the container, and still it wasn't completely secure.
The block chain shipping form solved this issue completely because you can never fake forms or hide when ownership of a container changes hands.
You could frankly do something similar for public records of landownership, which might not be important in the west where corruption is low, but could revolutionize the third world.
This is in early days in Maersk, who are an early adopter in shipping. Most in shipping have never heard of blockchains, and international standards like ICP600 have yet to mention it, banking trade departments not touching it yet (for things like LCs); Maersk are using internally and on small scale.
Very early ground. But a good starting point. Blockchains for shipping are pre-beta, but that doesn't mean testing isn't happing, yes.
Actually, this is better: http://iang.org/papers/triple_entry.html
My attempt at trying to explain it in one sentence: blockchain provides the technology to connect the double entry ledgers of two (or more) parties by recording the transactions between them, in a way that makes it possible to _prove_ the existence and validity of the transaction independently, without relying on either party's ledger.
It's like XML in the 90s or Web 2.0 in the 00s. Gotta use the technology to get the funding, regardless of actual suitability to task.
I think for most businesses, the real applicable portion of the blockchain was cryptographic signatures to solidify public ledger. That's about it. Kind of simple, but I think that's what people are paying attention to.
In effect, its less about the technology that BTC and/or cryptominers care about... and more about the boring part that seems to get people excited.
A lot of "blockchain" companies seem to be enabling peer-to-peer transactions for example. By centralizing all transactions to a particular server. There's this one company (I forget the name) which claims to be using Blockchain for exchanging Solar-credits between neighbors.
Having a ledger that is cryptographically reliable, even if centralized, is the main benefit. Also, one that can be fully automated is a big deal. I mean, that's all Visa or Mastercard really are: systems that describe when and where transactions have occurred around a centralized source of trust.
---------
EDIT: And yes, I know what a blockchain is in BTC circles. But I don't think a "Bitcoin Blockchain" is what people are talking about on typical marketing material. Just like "the cloud" has evolved to mean something new... "blockchain" seems to have been picked up by managers and/or venture capitalists to mean something totally different.
Because you just described how a bunch of businesses can make money in a way "Blockchain" cannot. From a financial feasibility perspective, it only demonstrates that a decentralized store of information is going to be innately unprofitable.
If you're old enough to remember the days before the internet, access to information beyond what your parents and peers could tell you was found in the public library. Books. I have always loved books for this very reason. They give me access to information beyond what my immediate community can (or will, when you think about it) tell me.
Of course, the problem with the internet (in its present form) is that the information is so unreliable. There are just too many conflicting interests. With the library, you at least have the librarian, who, being a sort-of gatekeeper, makes sure that the library contains only "good" books.
So, blockchain. What problem does it solve? It's a really good question. The blockchain is a distributed ledger, that is, basically, a key-value store. The keys are the bitcoin addresses, and the values are the bitcoins themselves. What is unique about it, and what makes it such an interesting program, is that the record of transactions is constantly being updated, because there is a bitcoin "reward" for running the program. So, even if everyone stopped trading bitcoins tomorrow, if they were just like, "nah, we're over it", the bitcoin network itself would still generate some new bitcoins, and the ledger would be updated. It's really a fantastic invention, when you think about it.
The value of a blockchain is that it adds a type of "reliability" on top of the internet. Because the ledger is public, anyone can inspect it. Because the network is distributed, all the miners running the program are, for want of a better phrase, "keeping eachother honest". When you look at a regular website (or book, or movie), it can be hard to tell if the information you're seeing is legitimate/honest. Whereas when you look at the record of the Bitcoin blockchain (for example), you can be fairly confident that those keys (addresses) were assigned those values (bitcoins) at that time.
To sum up, the problem that a blockchain solves is storing a reliable record of transactions, by harnessing the inherent self-interest of people to "make money" (ha ha).
It solves the problem of financial intermediation. The whole financial world has been built on financial intermediaries (banks and stock exchanges) to verify that you actually own things. Now you can point to a public record (the blockchain) for that, and no one can take it away from you (without your private key).
Blockchain would be excellent for verifying that a bag has been received and processed at a certain place.
Lots of the large GDSs such as Sabre/Amadeus/Travelport are actively looking into this.
For your example, cargo that changes hands between entities might be more interesting. US Airlines carry a fair amount of post office mail, for example.
But I suppose a DB could do that too. Maybe it's all hype ^^
How is the average person going to respond when they hear a new form of money is "trustless"? (I know what it really means in this context, but think about slope of adoption).
Just like today you probably don't ask your doctor the details about how they securely store your health records, nor do you talk to your bank about how they clear transactions.
So, we will see many concepts around trustless transaction verification in future, in the same way Napster has not been the final solution for P2P data share, but eventually led to BitTorrent.
IMHO the term "blockchain", in current and common use, actually means Merkle trees applied to time-series information that is shared between multiple computers over the internet, hence 'distributed database'.
Merkle trees are basically a way of saying "this precise set of information was used to generate this new precise set of information". Applied to successive timestamped data in a blockchain, this means that participants within a network can provide strong evidence for past state and state change - ie. they establish relative trust in the fundamental integrity of a shared and evolving database (the blockchain) based upon mutually agreed rules.
What is the real world advantage? Frankly, in most cases, there isn't one, but there are definitely disadvantages. Distributed databases are not new, and any other way of executing them is demonstrably more efficient, simple, etc.
The main potential advantage is to remove the ability of a central authority (or small colluding group of bad actors) to "change the rules" by "rolling back" the state of the system, halting the system, seizing control, removing actors, etc.
This property is perhaps at its most potentially impressive when its implications with respect to the conventional business environment, government and politics are considered for large global industries with deep pockets. Simply speaking, right now most regulations are issued in airy-fairy lawyer speak and cost a lot of money to skew/change/interpret/argue/take to court. This is the case for government regulation, case law and non-governmental agreements (industry codes of conduct, informal protocols, MOUs, contracts, SLAs, etc.).
The potential advantage of distributed systems with mutual trust (note that blockchains are only one implementation of this concept) are that all of this overhead can be removed, since an adequately cross-border (ie. cross-jurisdiction, or government-independent) network of an adequate size will be essentially immune from direct state actor intervention, and the rules for all parties remain clear.
To summarize: save lawyer money, ignore regulators, save time/hassle, remove any means of effective government intervention.
The reality, however, is that most industries are not able to agree on a clear set of interactions and rules under which to formalize their activities, and even if they could it would probably not be feasible to do so in all cases, therefore this semi-utopian objective cannot be reached, and in my professional opinion few corporate blockchain projects to date are delivering anything more than an inefficient means of distributing simple time series transaction data.
The big exception (ie. winning example for your last question) is Bitcoin, IMHO because it was the first, and because it quickly enabled safer and more cost effective means of recreational drug distribution, which essentially provided the initial market which funded its growth (now tending to speculation).
My prediction is that we will see a rise in smaller, private systems and asset types providing transaction services with different properties to present era cryptocurrencies and a smart layer will emerge to evaluate and utilize these based upon their objective properties and per-transaction requirements / risk models. Many of these systems will not be based on blockchain at all. Instead of de-facto bank or card systems, we will have a broad range of settlement mechanisms available. Instead of stock exchanges or large VCs, we will have a broad range of funding sources available. Access to capital will improve, business transparency will improve, government regulators and established financial services monopolies will slowly be disempowered.
Further reading - https://en.wikipedia.org/wiki/Merkle_tree + https://en.wikipedia.org/wiki/Blockchain + https://en.wikipedia.org/wiki/Time_series + https://en.wikipedia.org/wiki/Distributed_database + https://en.wikipedia.org/wiki/Bitcoin_network + https://en.wikipedia.org/wiki/Money + https://en.wikipedia.org/wiki/Cryptographic_hash_function
The problem is that whenever you have a public anything there will be spam. People will come along and they will fill the space up with scam ads and porn and hate speech. See: the internet.
To prevent the database from filling up with spam you need to find a way to make people "pay" for writing to the blockchain. Nothing is free and if you're going to consume resources (storage and cpu) from other's computers then they ought to get something in return. That's what "transaction fees" are and that's why they're measured in costs per byte -- the people who are writing to this public, decentralized, highly resilient database are being charged storage fees.
So blockchains are the solution to spam. And unlike pre-existing spam solutions a blockchain doesn't require storage/delivery nodes to "whitelist" certain certificates or for certificate authorities and CRLs and all that wackiness. This is kinda remarkable.
This innovation enables all sorts of new models of collaboration (read: business models). Just today I saw a very interesting proposal for an Uber-like system built on a blockchain. Everything -- requests for transport, driver availability, driver fees, driver certification, passenger and driver ratings -- gets stored in the blockchain. Anybody can read, anybody can write. Here's the database, here's the type of "transactions" (records) -- go crazy. The interesting thing about this system is that it doesn't rely on huge bureaucracy to protect the database. (And what are many business firms but database maintainers and protectors?) Write whatever you want -- write pics of your cats if you like -- but it's going to cost you.
(There's another element of what's commonly called blockchain which is "block verification." The problem here is that when everybody has a copy of the database how do you decide whose copy of the database is authoritative? This is the concept of consensus and note that it's only a problem if the peers in the database don't trust one another. It's actually not so interesting but it is largely an implementation detail of public blockchains with different strategies proof-of-work, proof-of-stake, electoral-validation etc.)
Edit: Not to mention, immutability means human errors cannot be recovered from - a wrongly worded contract cannot be restored.
Negotiations would be done by the "servicer" of the loan, who gets to charge a fee for this. There is also the "originator" of the loan (think the website where the person taking out the loan clicked on the affiliate link, but in meatspace). The "owner" of the loan is generally considered to be the person who currently puts up the (remaining) principal.
Each of these roles may or may not change hands as a result of a transaction, but generally only the risk role changes.
There is a large business "originating" loans, and over time during the credit cycle the business becomes originating ever more risky loans until the whole thing blows up and we get into panic, tighten lending standards (overtighten, generally), and go back into the business of loosening up.
I can't speak for the blockchain stuff but typically in financial software data is treated as immutable as well and errors are accommodated for by adding an additional correction record. I assume blockchain would enforce this practice, but then again, I don't know how it would distribute the amended contract.
Often, the people keeping the records of who owns what have become extremely powerful - it is my understanding that this technology has the potential to change that dynamic.
Go to the source code. This stuff is open-source and well documented. Once you start reading the code this stuff makes sense. Write your own blockchain. In fact it's not that hard and it'll make the basic concepts really sink in.
As I note elsewhere in this thread, you literally get that with Git, which was released in 2005 and based on previous work in DVCSes back to the late 1990s. And Merkle trees were invented in 1979.
The new thing in blockchains is the consensus mechanism. The best-known example uses as much electricity as Ireland.
One could call Git a "blockchain", but it'd pretty clearly be trying to synthesise a history for a neologism ...
I suspect "blockchain" products that survive will be tamper-evident transaction ledgers, and I already had one dev admit to me that his company's "distributed ledger technology" product was basically a simplified Git ...
(You still need to pay the ISPs of course and eventually the ISPs become too powerful and start demanding bigger and bigger fees and then the whole thing turns to shit. See: Bitcoin, the internet.)
But can't things like name of the person taking the loan can be changed? Obviously there is still a journaling entry recorded to track who and what changes were done. This is specially important because wrong names can and do affect credit ratings etc.
Nailing down this stuff is where you really appreciate a good architect.
You don't have to replace 100% of what something does to disrupt it
A great use case is in EHR. Your doctor, your insurance company, and all their BAAs participate in private block chain that controls access to your records. Any time a record is accessed, the chain is updated.
Nodes can participate in establishing consensus without having the ability to decrypt the record.