Occasionally I get asked questions like “What is this blockchain thing?”, “What is the difference between a consortium and a public blockchain?”, “Which should I choose for my project?” and so on.
As the hype runs amok, such questions come more and more from non-technical people, managers and decision makers, and while I found that the narrative of the immutable distributed ledger is good enough to answer question one, it gets harder to communicate the answers to question two and three. Even simple graphical decision charts like the one published by Wüst and Gervais seem to be to technical, too abstract.
So, when driving home one day after another meeting full of cognitive confusion, it came to me that there must be another way to think and communicate about blockchain, with all the techno-speak and blockchain-will-save-us-all hype removed.
What is the core promise of blockchain? Trust.
What does (hopefully) every manager understand? Trust.
How about treating blockchain as a trust model, then?
Since the Soccer World Championship 2018 has just begun, let’s play three rounds of Blockchain Soccer as an example of a trust model.
First, let’s define the stakeholders in the soccer model:
- The Teams. There are 2..n teams in the tournament and they all have a strong incentive to win the tournament, because they like soccer, because it creates visibility and fame, and because it generates revenue through purses, merchandise and sponsoring contracts.
Since the stakes are so high, some teams could be tempted to bend the rules of the game a bit to their favor.
- The Fans. They want their teams to win, but also to see good and fair soccer for their entertainment. Cheating teams ruin the experience in the long run.
- The Sponsors. The sponsors want to see good and fair soccer because they pay the teams to gain visibility among the fans. Cheating teams will ultimately hurt the credibility of the sponsors.
- TV Stations. TV stations pay considerable sums for the broadcast licenses, so they are also not interested in cheating teams, since cheating will hurt their ability to sell expensive advertisement time during highly visible soccer tournaments.
As you can see, all stakeholders share a common goal of playing fair soccer by the rules, however individual teams might also see a short term incentive in bending the rules to their favor. There is some mistrust between the teams and also between the teams and the other stakeholders.
Round 1: Trust through an impartial referee
The usual way to instill trust on a fair game between all participants of a soccer tournament is through a impartial referee or rather a team of referees.
The referee guarantees that the teams play by the rules. If he recognizes a violation of a rule, he can punish the violator so that no team gains any advantage by cheating.
So the existence of the referee establishes a trust boundary between the teams and the other stakeholders, indicated by the blue line in figure Round 1.
That’s the theory and it works quite well for some time now, but it has some shortcomings:
- The referee might not see every violation of the rules
- The referee might not be truly impartial
- The referee is part of the game but no player so she might create unwanted interference
So assume that there would be a way to replace the referee by a advanced surveillance technology. Let’s call it the mechanical referee. The mecanical referee records all movement on the playing field, every contact between the ball and the players, every contact between the players. Everything. Its advanced AI can act like a human referee, except that it is omniscient and does not need a physical presence on the lawn. Also, it keeps a record of all past actions that can be used to analyze the game ex-post.
While such a mechanical referee could be seen to cure some of the issues with a human referee mentioned above, one issue remains: It might not be truly impartial. Somebody has to operate it, and whoever operates it can alter the database where all the actions are recorded in a central place. To address this issue, this central database must be safeguarded, e.g. by audit trails, tight access control, and digital signatures. Again, some impartial third party is needed to create trust, we just moved the referee from the lawn to the datacenter.
This is how most (if not all) central database systems that carry vital data are safeguarded today, be it in banks, insurance companies, or healthcare. At least I hope so…
Enter the blockchain and on to the second round.
Round 2: Trust by blockchain
As said before, the teams share a common goal of playing fair in the long run, but might be tempted to abandon this for gaining short-term advantages. This means that there is always some level of mistrust between the teams about sticking to the common goal since they are competitors for limited resources. A team might try to influence the operator of the mechanical referee to alter data in their favor.
But if the teams agree on dedicating some compute resources, the central database of the mechanical referee could be replaced by a blockchain shared between all teams.
The more teams participate, the lower the probability that one, or a group of teams gain enough majority to alter data on the blockchain. And the fact that they are still competitors, basically mistrusting each other lowers the chance of cheating even further.
This setup is shown in the figure for Round 2. Now there are local trust boundaries for every team, indicated by the red lines, but exactly this level of mistrust between the teams, combined with a blockchain generates the same trust boundary towards the other stakeholders as the central referee in Round 1 did. Except that the trust is now shared, or decentralized, between all teams alike.
And this is the trust model of a consortium blockchain!
Can we do better?
Round 3: Going public
With our consortium blockchain in place for some time now, rumours are that, against all odds, some cartel and camaraderie has been built up between enough teams to make manipulating the mechanical referee a possibility again.
More trust is needed! Time to go public as shown in figure Round 3.
By extending the blockchain to the rest of the stakeholders, additional trust is gained. First, the blockchain is now spread over many more participants. It becomes even more decentralized.
Second, there is some additional distrust introduced to the system, e.g. between the sponsors and the teams — and even where there is no immediate distrust, e.g. between the TV stations and the fans, the sheer lack of other common interests except soccer, minimizes the chance that a large enough group of participants cooperates to make cheating possible again.
The trust boundary has now been expanded even further by incorporating more public distrust (and mutual ignorance) into the blockchain.
And this is the trust model of a public blockchain!
And if you now think back to Round 1, you’ll immediately see that a private blockchain makes no sense in terms of a trust model. It generates the same level of trust than a conventional central database, only with more operational problems.
Analogies to replace mathematics are always a dangerous path to follow. Still I hope this helps to demystify the blockchain hype a bit more and to enable decision makers to see beyond the technology and to think in trust models, where they might feel more comfortable.
Yet, some interesting questions remain. How to incentivate the participants in a blockchain? Compute power costs money, and while it might be comparably easy to justify this in a consortium blockchain model, it might get more difficult in a public one. Cryptocurrencies, like Bitcoin, sidestep that problem by creating new “money” out of thin air, but that is obviously no general solution.
And can trust generation, and individual contribution to trust generation, in a blockchain system be measured in some way?
Questions worth answering methinks.