Bitcoin security model: trust by computation

http://radar.oreilly.com/2014/02/bitcoin-security-model-trust-by-computation.html

by Garbage • 12 years ago

51 25 12 years ago

radar.oreilly.com

patio11• 12 years ago

This is a neat trick, in that it defines Bitcoin's security model such that all of the successful attacks are out of scope. Meanwhile, the central actors Bitcoin doesn't have who received the trust that Bitcoin doesn't require ran into a smidgen of trouble running the bank that Bitcoin makes obsolete and now their depositors that Bitcoin doesn't acknowledge are praying that the government that Bitcoin undermines will hit them with the legal process that Bitcoin structurally avoids.

GotAnyMegadeth• 12 years ago

Difficult sentence, but I got there in the end :-)

torbjorn2• 12 years ago

The Gox situation is unfortunate but I think it stems from the weaknesses and corruptability of centralized systems (in this case the Mt Gox Bitcoin Exchange) rather than flaws within Bitcoin.

Decentralized exchanges are in the works that should prevent similar situations in the future.

hershel• 12 years ago

assuming exchanging bitcoin with usd becomes illegal with big enough penalties for both sides, what value would a decentralized exchange have, except for illegal businesses?

alwillis2• 12 years ago

The Gox situation is unfortunate but I think it stems from the weaknesses and corruptability of centralized systems (in this case the Mt Gox Bitcoin Exchange) rather than flaws within Bitcoin.

Mt. Gox has been a poorly run company as long as it has existed; it didn't help that its custom bitcoin wallet implementation didn't take transaction malleability into account, even though it's been known about since 2011: https://bitcoinfoundation.org/blog/?p=418

Remember that Mt. Gox has about 20% (down from 80% back in the day) of Bitcoin's trade volume; the other major exchanges properly addressed malleability and didn't have the same problems.

gwern2• 12 years ago

Meanwhile, the central actors Bitcoin doesn't have who received the trust that Bitcoin doesn't require ran into a smidgen of trouble running the bank that Bitcoin makes obsolete

Mtgox is not a centralized bank in the Bitcoin economy, has steadily been deprecated due to its ongoing problems, and long ago lost even a bare majority of Bitcoin/fiat turnover as people moved away from it (just as a distributed currency lets them easily do). Its problems are now between it and the customers who decided to bet on it in trying to take advantage of the famous Mtgox risk premium.

Bitcoin: working as designed.

sheldonth• 12 years ago

Patio11, I have much respect for your work and _love_ reading your email digests, but your comments of late about Bitcoin have such a decidedly negative slant that they become hard to digest at face value.

Yes of course businesses built around the Bitcoin protocol will melt down occasionally, as many businesses do within the course of operation -- but don't you see how your attempt to conflate Bitcoin businesses (Exchanges, specifically) with a discussion about the new model of trust employed by the p2p network is probably a bit deceptive?

Yes, Bitcoin is scarce and therefore commands a market price. Commanding a market price has some qualities that suck. One of them is that you have to deposit assets in a market to keep an order book to discover a price. This is fallible. But it has nothing to do with how we (the human race) have discovered a way to create a scarce and transferrable peer to peer asset over the internet. Perhaps you should stop and consider how you can personally deconstruct some of the negative bias you have against Bitcoin in your public discussions of such.

Much love.

[Edited for grammer]

Hieroglyph• 12 years ago

Commanding a market price has some qualities that suck. One of them is that you have to deposit assets in a market to keep an order book to discover a price.

You don't _have_ to trust a counterparty with your money for price discovery. Take Counterparty for example: a P2P derivatives and stock market using Bitcoin as the transport layer [ǂ]. It enables trustless betting on anything from asset prices, to weather events, to Superbowl results [ɵ]. Two companies have already issued publicly tradeable shares [ɸ].

Data feeds are a public, competitive market:

    Order: give 55.0 XCP for 0.5 BTC in 11 blocks, with a provided fee of 0.0001 BTC and a required fee of 0.005 BTC (bcf02fb66565d984a136fc55e1085f5cc782e8630dcfdbe9bbddb3312beb8f2c) [valid]

    Broadcast: ‘Block Hash (0000000000000…cdbe902b4698e) Even/Odd’ = 2.0 from 15cdAQmmBrz1BEVtipaQ1dVHtTwmfcxzw5 at 2014-02-20T20:00:03-08:00 with a fee of 0.1% (7bd292f02c41740150b2ed2e7cf739566c900092e892e4a3a5d7dd142c00df8b) [valid]

    Broadcast: ‘CoinDesk BPI USD’ = 555.26 from 1CeQHd59TFKWQzsWYDXc9NDX2ooMSRpiqi at 2014-02-20T22:00:03-08:00 with a fee of 0.1% (65c7162d69664008e4e0fe3be9122fb02c3e98901cd975d6d179e85a945fd916) [valid]

[ǂ]: https://counterparty.co

[ɵ]: http://blockscan.com/tx.aspx?q=3155

[ɸ]: http://blockscan.com/assetInfo.aspx?q=MPTSTOCK

[ɸ]: http://blockscan.com/assetInfo.aspx?q=SFMSTOCK

unclebucknasty2• 12 years ago

I think what patio11 is saying (and I agree) is that too much is being made of the ideals behind the Bitcoin protocol and paradigm, but it doesn't really translate when met with the hard test of practical reality.

I noted your argument, along with many others on the GP's sub-thread which make this very distinction between Bitcoin as a concept and Bitcoin as an actual thing. In so doing, you suggest that failure to make the distinction is misleading and you align yourselves with the idealistic praise heaped on by the original article.

But, I would agree with patio11's insinuation: i.e. that the real fallacy is in this praise. It's misleading, because at some point, the protocol must be implemented in the real world, consisting of exchanges, wallet services, payment services, etc. These are all points where bad actors can attach. To say, "well, that's not the Bitcoin protocol" is to point out a meaningless distinction. When you find a way around these implementation limitations, then it matters.

TL;DR: as implemented, Bitcoin in practice solves exactly none of the problems espoused by the article. We need look no further than real-world events to see this.

mikeyouse• 12 years ago

This is a good point, but I had difficulty reading it without adding some parentheticals;

    This is a neat trick, in that it defines Bitcoin's
    security model such that all of the successful attacks
    are out of scope.

    Meanwhile, the central actors (which Bitcoin doesn't
    have) who have received the trust (that Bitcoin doesn't
    require) ran into a smidgen of trouble running the
    bank (that Bitcoin makes obsolete) and now their
    depositors (that Bitcoin doesn't acknowledge) are
    praying that the government (that Bitcoin undermines)
    will hit them with the legal process (that Bitcoin
    structurally avoids).

oleganza• 12 years ago

Consider that people freely decided to keep money in MtGox. The whole currency is unaffected. But when your central bank decides something (e.g. to issue more currency or do a haircut), then suddenly all banks and all holders are affected. Example of MtGox is an example why Bitcoin is important and actually works as promised: one company misbehaves, others not affected.

applecore• 12 years ago

There's bitcoin (the unit of measurement), Bitcoin (the software, system, and distributed consensus network), and "Bitcoin" (the people, community, and economy that have emerged around Bitcoin in the past five years). Your argument conflates Bitcoin and "Bitcoin" so seamlessly that the reader may not even know there's a distinction.

They seem to be a single identity to you, and the differences appear to become lost. The misunderstandings and errors of logic in your argument come from treating two distinct concepts as if they were one. For instance, the term "security model" has a distinct meaning when used in the context of a distributed computing system, but that doesn't seem to matter to you.

drewblaisdell• 12 years ago

They seem to be a single identity to you, and the differences appear to become lost.

This is a good description of patio11's Bitcoin arguments. He ignores the problems Bitcoin does solve and focuses on the aspects of the community that are broken and still need work.

Gates called Bitcoin a techno tour de force. It has potential and solves some real problems, but it has a long road before becoming anything like a usable currency. Patrick tweeted about HN comments of this sort last week and it sums up exactly how I feel about his Bitcoin comments: https://twitter.com/patio11/status/435315147427430400

roymurdock• 12 years ago

Interesting comparison between Bitcoin and the internet. The Umlaut published an article outlining a similar theory: (http://theumlaut.com/2014/01/08/bitcoin-internet-of-money/) using the same terminology (transport layer) and describing Bitcoins as the "internet of money".

I noticed an interesting comment on the O'Reilly article: "The second problem is that we have yet to expand bitcoin beyond the control of virtual assets. Virtual assets do not physically exist and therefore it is the ability to represent these assets within multiple computers across the network which give bitcoin its decentralized nature. These virtual assets also have to have real world value to everyone, without value, there exists no economic incentive (block rewards and transaction fees) for securing the network through computation."

This analysis of Bitcoin as the "internet of money/finance" is kind of reminiscent of the "internet of things" movement, that, according to inept marketing departments, is guaranteed to revolutionize our lives in short order.

Do we have the need/infrastructure in place for an "internet of money" or an "internet of things" to emerge? Do these terms even hold any weight? Would love to hear any opinions.

tlrobinson2• 12 years ago

While there’s an element of marketing-speak to this, I think it’s instructive to try to apply the idea of “Internet of [whatever]” retroactively. The internet originally was essentially the “Internet of Communication” (email) and “Internet of Information” (FTP, then HTTP). Advancements in technology enable new use-cases over time.

* Internet of Commerce: mainstream adoption of the internet

* Internet of Entertainment: broadband internet

* Internet of Things: ubiquitous network-enabled electronics, with size and cost reductions driven by the smartphone revolution

* Internet of Money: Bitcoin

A corollary to “Software is Eating the World” could be “The Internet is Eating the World”. Any technology that enables one of these new “Internets” should be watched closely.

higherpurpose• 12 years ago

Here’s the most important effect of this new trust model of trust-by-computation: no one actor is trusted, and no one needs to be trusted. There is no central authority or trusted third party in a distributed consensus network. That

That should be the future of the Internet, where we don't have to trust whether someone is a "good actor" or a "bad actor". This is also what is needed so we don't have "balkanized Internet" because we can't trust countries like US or UK to not spy on the rest of us anymore and get our data. If the network is built in a way that no one can be trusted, then the network can remain truly global again.

That's why I'm glad to see projects such as the Turing complete Ethereum picking up steam, and I'm hoping they succeed. I'm starting to believe more and more that the future of the Internet is going to be "Bitcoin-inspired" in a major way, and it will bring us a lot more distributed trustless applications than exist today on the current Internet, making it a lot harder for "bad actors" to exist in the network, be that for-profit hackers, "cyber terrorists" or even malicious governments.

It's also what will save us from the evil of certificate authorities, the OpenSSL monster, and other such patched up security tools that are meant to protect us right now, yet mostly fail to do so.

platz• 12 years ago

I think Andreas Antonopoulos said in a video that it takes about half a billion dollars to fool the blockchain for 10 minutes. Even though that's a lot, if certain actors (state or otherwise) got interested enough, wouldn't they manipulate the chain at critical moments for their advantage?

alwillis2• 12 years ago

I think Andreas Antonopoulos said in a video that it takes about half a billion dollars to fool the blockchain for 10 minutes. Even though that's a lot, if certain actors (state or otherwise) got interested enough, wouldn't they manipulate the chain at critical moments for their advantage?

That's out-of-date now; the total computation power of the Bitcoin network is 25,824,939.77 GH/s. It would probably take a lot more than $500 million now.

gwern2• 12 years ago

It might take $500m to buy the hardware at current prices, but I'm sure you could bribe the ghash.io admins with well under $500m to hire their pool for a few minutes.

thisiswrong• 12 years ago

Andreas is an impressive character. Only by listening to him speak will you understand the passion [1] and potential this guy has. He is well versed and articulate and has a background in distributed systems. Like some of us here he once worked for thieves & sociopaths (banking/defense), and dropped it all to dedicate himself to bitcoin. He has a strong presence in the bitcoin communities and visits many meetups.

I strongly recommend listening to some of the early episodes of him, Stephanie, and Adam on 'Lets Talk Bitcoin'. There's some deep topic discussions and interesting projections - most which turned out to be true.

[1] https://soundcloud.com/mindtomatter/e72-powerful-perspective...

Bitcoin security model: trust by computation

Related Stories