All field notes

Blockchain · Retrospective

Masternodes already bond collateral: Trivechain's model beside the beacon chain

Two pedestals with locked glass boxes of coins, and a small gavel between them.
Concept illustration · AI-generated

Ethereum's beacon chain started on 1 December, run by validators who each deposited 32 ETH. On Trivechain, where I lead development, every masternode already bonds 10,000 TRVC to take part in the network, and I wrote the first version of our masternode registration guide in September 2019.

Putting the two side by side shows that they share a word and use it differently.

What the collateral is

Trivechain's masternode code descends from Dash. To register a masternode, an owner sends exactly 10,000 TRVC to a single output, waits for 15 confirmations, and submits a registration transaction. That transaction names the collateral output, an owner key, an operator key, a voting key and a payout address. The guide points out that the collateral wallet can stay offline, in cold storage, when it signs the registration message.

The collateral never leaves the owner's wallet. It remains an ordinary unspent output. If the owner spends it, the core code removes that masternode from the list in the same block. Leaving is always possible and takes one transaction.

On the beacon chain, the 32 ETH goes into a deposit contract on Ethereum, and for now there is no way back: withdrawals are left to a later upgrade. Each deposit commits to withdrawal credentials, tied to a key kept apart from the signing key that the validator uses once per epoch, every 6.4 minutes.

What can be penalised

A Trivechain masternode is held to a Proof of Service score. One that fails to do its share of the network's work collects penalty points, and past a threshold it is banned and stops receiving its turn at the masternode share of block rewards, currently 35% of each block reward. The operator can revive it once it is fixed. The 10,000 TRVC is never touched. The worst a masternode stands to lose is income.

A beacon chain validator can be slashed. Proposing two blocks for the same slot, or signing two attestations that contradict each other, destroys part of the deposit and forces the validator out. A further penalty grows with the number of validators slashed in the same period, so a failure shared by many costs each of them more than a failure alone. Validators that are offline lose smaller amounts, and more if the chain stops finalising.

The difference follows from the role. On Trivechain, miners produce blocks with proof of work, and masternodes provide services alongside them. On the beacon chain, validators produce and finalise the blocks. A validator that signs two histories is attacking consensus, so the bond has to be forfeitable. A masternode that goes offline mostly degrades a service, and losing its payments is a proportionate cost.

Who votes, and on what

Both systems talk about voting, about different things.

A masternode's voting key votes on TriveGovernance proposals, which cost 50 TRVC to submit, one masternode to one vote. Sentinel, which runs alongside the core software on each masternode, handles the governance objects. Those votes decide what the network's treasury funds.

Beacon chain validators vote in every epoch. Their attestations say which block they see as the head of the chain and which checkpoints should be final. These are votes about history, not about money or direction. Decisions about changing Ethereum's protocol are still made off the chain.

What I take from the comparison

Separate keys by job. Both designs keep the key that runs on an always-on server away from the key that controls the money. On Trivechain the owner, operator and voting roles can each have their own key, and rewards go to a separate payout address. On the beacon chain the signing key and the withdrawal key are separate from the start. This is the part I would copy into any system that holds value.

Bonded is not the same as forfeitable. Masternode collateral limits how many identities one person can afford to run, which makes it an entry ticket. A slashable deposit does that and also makes a specific kind of misbehaviour expensive. Which one a network needs depends on what an operator could break.

Liquidity is part of the commitment. A masternode owner can leave in one transaction. A beacon chain depositor cannot leave yet. Someone comparing returns between the two models is also comparing very different commitments.

I maintain a chain built on the older model, so I know its strengths and the questions it leaves open. Watching a large network start the newer one this month is a useful way to see which parts of each design are choices and which are necessities.