How we think about decentralized systems
“Decentralization" is a popular term used to bracket a set of interrelated technologies including blockchain, crypto tokens, NFTs. It’s also a fuzzy term that has been used in subtly different ways in different disciplines throughout history, so the question “what is a decentralized system", or, more to the point, “is this particular system really decentralized" can be quite difficult to answer unless we have a clear definition by which we can measure ourselves.
Decentralization and control
Decentralization is ultimately about control, and the widely used political economy definition usually requires the devolution of some measure of control from a central actor to multiple actors. According to Wikipedia, decentralization looks like this.
But it’s easy to see that the decentralized system maintains one major disadvantage of the centralized system: remove the central node and the system falls apart. Indeed, the graph on the right could simply be rearranged to depict the supposedly decentralized system as a multi-level hierarchy.
Even if some project clearly have this definition in mind, this is not how we conceive decentralized system.
Decentralization and mutuality
For us, decentralized systems are systems of mutual control. No single actor has privileged rights, to add or remove data, to make unilateral decisions, to disconnect or replace other actors. No single actor has information or decision making advantages. No single actor controls information flow.
A single truth without a single source
In computer science, a single source of truth is any actor that, if there is disagreement about the state of the system, can overrule all others. Notably, the definition of “truth” in this context is not “what is true” but “what is held to be true”. So a controlling actor could perfectly well seed false information and make all others adopt it.
Decentralized systems are designed to avoid this scenario.
This means in practice that decentralized systems allow every actor to propose new information to be included, but all other actors will test this new information for consistency before they accept it. And only new information that is consistent for all actors will be accepted.
Depending on the setup this mutual up-front control of incoming information can be quite resource intensive — a common critique of decentralized systems.
But the key to understanding the economic motivation for decentralized systems is that this process shifts the cost of dispute resolution from after the fact (often a long time after the fact when dispute resolution becomes a very costly endeavor) to before the fact.
The result is a system that is known and accepted by all actors to reflect consistent information, with no single actor having the privilege to control information flow.
Decentralized systems for decentralized scenarios
Clearly, this kind of a setup is not a catch-all for all scenarios. To discover scenarios, we have to find situations where these premises hold true: we don’t want a single actor to be able to overrule everyone else, and we want information flow to be non-excludable.
Luckily, over the last years value creation has gradually shifted from orchestrating production (inside firms) to orchestrating exchange (inbetween firms).
The current solution to this orchestration problem, platforms, tend to be very lucrative for the platform owner but disadvantageous for everyone else because they ultimately create a hierarchical structure for orchestrating exchange: the platform owner retains control over access and dispute resolution and critically, the ultimate information advantage.
To alleviate these problems we should be thinking about decentralized systems.
