The CAP Nemesis: Ensuring Fault-Tolerant Coordination
Distributed systems can either be consistent (everyone has the same data) or available (each node responds with its latest copy). Having an anonymous and decentralized environment on top of that makes it virtually impossible for nodes to trust each other's messages.
When the Bitcoin whitepaper was introduced in 2008, it drew on decades of cumulative research in computer science, culminating in a resilient “peer-to-peer” solution—an incorruptible, trustless network where nodes could collectively agree on what is true with a high degree of confidence.
Leslie Lamport’s work on the Byzantine Generals Problem was central to overcoming CAP limitations and establishing fault-tolerant coordination between nodes under adversarial attack.
Lamport’s 1982 Thought Experiment
"A reliable computer system must be able to cope with the failure of one or more of its components." Whatever happens to individual nodes and whatever behavior they may exhibit, the network has to be able to continue operating unaffected.
Leslie was particularly interested in the type of failures inflicted by nodes who propagate conflicting information through the system. And to better understand how to cope with the situation, he represented it abstractly as the Byzantine Generals Problem.
Imagine there's a group of Byzantine army generals encamped with their troops around an enemy town. They must agree on a common attack plan and the only way to communicate with each other is by messenger. One or more may be traitors who try to confuse everyone. The goal is to find an algorithm that ensures the loyal generals reach agreement.
Simple Plan
After observing the enemy the generals have to align on a plan of actions and attack. So there has to be an algorithm ensuring the following conditions:
{1} All loyal generals decide upon the same plan. This is the condition the algorithm has to guarantee regardless of what the traitors do.
{2} Any cohort of traitors cannot cause the loyal generals to accept the wrong plan.
So what Leslie proved was that for a solution to the Byzantine Generals Problem using oral messages to cope with m traitors, there must be at least 3m + 1 generals (where m is the number of traitors). For example, suppose m = 1. Then the whole system has (3 x 1) + 1 = 4 generals with one betrayer. In case of m = 0, the probability of a successful communication is 100% lol.
If the generals can send only oral messages (= contents are entirely under the sender’s control), then no solution can succeed unless > 2/3 of the generals are oriented towards success of the mission (= loyal generals in strict terminology). A traitorous sender can transmit any message in oral form and with only three generals, no solution can succeed in the presence of a single traitor.
Signed Messages > Oral Messages
Introducing incorruptible signatures to the network now makes requiring four generals to cope with one traitor obsolete. Moreover, in this case a solution for 3 generals does indeed exist. The following assumptions enable the network to succeed with m traitors for any number of generals:
{1} A loyal general's signature cannot be forged, and any alteration of the contents of his signed messages can be detected.
{2} Anyone can verify the authenticity of a general's signature.
The commander dispatches a signed order to each of his lieutenants. They in turn add their signatures to that order and send it to the other lieutenants, who add their signatures and send it to others, etc.
"This means that a lieutenant must effectively receive one signed message, make several copies of it, and sign and send those copies. It does not matter how these copies are obtained; a single message might be photocopied, or else each message might consist of a stack of identical messages which are signed and distributed as required."
Sounds familiar, right?
BGP <> Blockchain: Implementing a Reliable Computer System in Electronic Cash Exchange
The problem of coordinating between nodes is similar to the classic scenario of army generals planning a coordinated attack. The idea behind the consensus mechanism is to connect a network of computers so that it functions, in a sense, as a large supercomputer that continues operating (= mission success) even if some nodes fail.
Bitcoin implements a Proof-of-Work consensus mechanism to create an immutable chain of data records containing timestamps and digital signatures of identities participating in the network. "We define an electronic coin as a chain of digital signatures." So the coin (BTC) is sent to someone by digitally signing a hash of the previous transaction and the public key of the next owner, and adding them to the end of the coin. A payee can verify the signatures to confirm the chain of ownership, thus eliminating the need for an overseeing authority.
However, the double-spending problem could be an issue here. So transactions must be publicly announced to the network, and participants need a system to agree on a single history of the order in which coins were received. That’s where Satoshi was drawing on the BGP ideas: the payee needs proof that, at the time of each transaction, the majority of nodes agreed it was the first received (= accept something to be true with a high degree of confidence).
Validators and Cypherpunks
Proof of work also solves the problem of determining representation in majority decision-making. In Leslie's experiment it was somewhat implied that everyone agreed on who the generals were. And in theory, the enemy could be 99% of the generals, which would undermine the whole computer system. Cypherpunks were against having a list of known privileged delegates to run the network, so Satoshi tied participation to economic resources. Each identity solves a math problem that requires a certain amount of energy, and if five identities run it, you have five identities (effectively weighted by their work) running the network.
If voting were tied to addresses (one address equals one vote), an attacker could game the system by creating many addresses. Proof of work instead links voting power to computation. One CPU = one vote. The network’s decision converges to the longest chain, the one with the most cumulative computational work. If honest nodes hold the majority of total hash power, the longest chain will advance fastest and outpace any forks (any divergence in the main chain’s history or rules). Attackers would need to redo the work for the target block and all the blocks that came before to overtake the honest chain.
The Sentient economy uses token emissions to reward valuable contributions and energize the network. The DAO controls the emission rate and distribution, ensuring the ecosystem remains healthy and adaptive.
But how exactly do Polygon, Arbitrum, and Celo L2s relate to it?
A
