Bitcoin as infrastructure [iii]

August 9, 2014 by Lucian Mogosanu

Also read: Part I, Part II.

Part III: Bitcoin as infrastructure

Now that I have described "the bigger picture", I can finally focus on the subject that I've set out to dissect. I promised that I wouldn't get into details pertaining to finance and economics; however, I feel that I must provide some background, especially that I myself wasn't very aware of Bitcoin at the time when it appeared.

The idea of "electronic money" isn't by any means new; it had been an interesting research subject in the field of distributed systems before 2009, the year when the anonymous Satoshi Nakamoto published his Bitcoin white paper. The main idea behind virtual currencies is that currency is in itself a token whose value is (theoretically) determined by the value of various resources in the market1. In other words, money has no intrinsic value and it's used only to quantify the value of useful stuff: for example I cultivate tomatoes and sell a bunch of them for $x$ coins. Then I buy some other stuff I need, say, computer games (since I don't produce computer games), for $y$ coins, and since I only possess $x$ coins, I can only buy computer games for $y leq x$. Thus coins are a mediator, and a useful one indeed, since I don't know (and don't care) whether computer game producers need or want to acquire tomatoes or not.

So virtual currency is, despite what some people might think, the ideal form of currency, if we assume that energy spent on running computers has less value than the costs of producing paper and/or metal tokens. The open problems before 2009 remained: who decides how many monetary units go into the market, and who intermediates transactions?

The obvious answer to that would be the state and banks respectively. However, we should note that banks are institutions traditionally used for depositing and withdrawing money, plus credits and the such. Allowing them to intermediate each single transaction would turn them into a single point of failure, and it's stupid, to say the least, to assume that the system is infallible. The same could be argued about the first problem: the state could in theory decide how much money gets to be produced, e.g. instead of tying it to some scarce resource, but we can't expect the state to make decisions that will satisfy everyone. Thus Satoshi proposed an alternative design which is decentralized, i.e. it eliminates the state and banks from the equation, and which eliminates some technical problems, i.e. double-spending, by design.

The way it does that is still poorly understood by most people, therefore I will attempt (and hopefully will not fail) to explain it2: assuming that the network contains money, a number $n > 0$ of accounts are open at a given time and that the sums of money in all accounts are represented by a set $A = {a_1, a_2, dots, a_n}$, we could in theory create a transaction $t_{ij}$ from $a_i$ to $a_j$, $a_i, a_j in A$; thus $i$ transfers $x$ units to $j$ and $a_i$ and $a_j$ get updated3.

Assuming that all the information above is accurate, the question is, how do we prove that the transaction was made? i.e. that $a_i$ loses $x$ units and $a_j$ gains the same amount. The natural solution proposed by Satoshi involves keeping a public database $T$ of transactions, which records all the transactions from the beginning of the world to the current moment. Thus, if $i$ updates his copy of the database and $j$ does the same thing, and $k$ and so on do the same, then eventually the majority of the network will agree on the transaction; as a consequence, all transactions are public. This is not entirely dissimilar to the way DNS works, the main difference being that DNS is structured hierarchically, while Bitcoin is implicitly peer-to-peer.

Having said all the above, the Bitcoin database, known as the "blockchain"4, didn't start as I described it earlier. There was no money in the system, which Bitcoin solves by introducing a "proof-of-work" framework: to create money, one must "mine" it by solving a cryptography problem, e.g. finding the transaction block which encodes to a particular hash. The problem's difficulty increases as the database grows, and thus virtual coins, i.e. bitcoins, become more and more scarce as the monetary mass grows. Miners are however central to Bitcoin by providing transaction validation (the "agreement" mentioned above) and are incentivized through transaction fees.

I won't discuss the model's validity and soundness5 here, but it's safe to say that the community of Bitcoin enthusiasts are testing it and discussing its problems since its beginnings, which doesn't automagically make the system perfect, but it shows that it has worked in practice until now. I will leave the technical debates to people more knowledgeable than myself.

It should be noted that while Bitcoin rests on top of the Internet's infrastructure, it doesn't specifically require it: in theory, one could validate transactions using pen and paper, although computers and the Internet make it practical. It should also be noted that Bitcoin as I've described it is more than a currency in the usual sense: it is in fact a protocol for distributed transaction validation through voting.

In other words alternative blockchains can be created to use the same infrastructure for various other applications, e.g. voting, file storage, social networks and so on, and so forth. The possibilities are quite hard to imagine, but we could begin with every piece of centralized software that would naturally run in a decentralized fashion6 using a blockchain. Granted, this would probably break the Web as we know it, but it would also break the feudal model of the current Internet.

In the end this could prove to be an impractical dream, a dream which I'll remain skeptical of just for the sake of pragmatism. But this aside, I can't help but recognize the potential of this thing, since it's not so different from how telephones looked in the '60s, personal computers in the '80s or the Internet in the '90s. And to prove that I'm not the only crazy person thinking about this, and more, that I'm not the first, nor the last, getting the idea of Bitcoin-as-infrastructure, I will provide some examples of interesting applications based on the infrastructure provided by Bitcoin before drawing a conclusion to my essay.

Next: Examples and conclusion


  1. Rigorously speaking it is in fact established by the agents, based on the value of other resources. E.g. if the demand for apples is high but the supply is low, then the price-per-unit of apples in some arbitrary currency will necessarily rise; on the other hand, if no one wants to use said currency for exchanging apples, then the (non-intrinsic) "value" of said currency will plummet. As far as economics go, this is mostly empirical evidence, although the reasoning itself is sound. 

  2. Not unlike many Haskell programmers attempt to explain monads and fail. But it's the attempt that matters the most, as it crystallizes the idea in the head of the storyteller. 

  3. This is from an imperative rather than a formal mathematical point of view, since identifiers never "get updated" in mathematics, rather they are different at different (explicitly given) points in time $t$. The two views are equivalent though, and one can be expressed using (subsets of) the other. 

  4. Due to it being a chain of signed cryptographic hashes of transactions. 

  5. Here's an interesting question: what would happen if a single entity managed to achieve more than half of the network's mining power? And what if that entity happened to be the state? I guess we're back to square one then. 

  6. Facebook, most Google services and a plethora of other web applications that kill their users' privacy for the sheer purpose of serving better ads. 

Filed under: asphalt.
RSS 2.0 feed. Comment. Send trackback.

One Response to “Bitcoin as infrastructure [iii]”

  1. [...] coming from Bitcoin11. Although the idea was conceived specially for implementing a new form of representing money, its uses may theoretically go beyond that, into other distributed systems and [...]

Leave a Reply