Perun: Virtual Payment Hubs over Cryptocurrencies

The five-minute read

One action, walked through

1. Each user opens a ledger channel with the hub on the underlying chain.
2. The three parties agree to a virtual channel with a capacity and a timeout.
3. The two users update that virtual channel off-chain for each payment.
4. They close cooperatively, and the hub's ledger channels absorb the net.
5. If someone cheats or disappears, the contracts on the ledger are the enforcement path, inside the timeout.

The argument, unpacked

What has to be true

• The underlying chain executes the contracts the construction uses.
• The hub and the users can be forced to settle before timeouts.
• Virtual-channel capacity was actually locked. An unfunded virtual channel is an IOU outside the paper.
• The adversary model is the one in the proofs, which includes a malicious hub. Read which theorem covers which party.

What happened after the paper

Perun is the citation for virtual channels through a hub, next to Lightning and Sprites. A payment system that says 'we are like Lightning but the hub is offline' is pointing at this idea, and still owes the dispute contract.

What to check before you use the idea

• Does a normal payment require the hub to be online?
• Which chain holds the dispute contract?
• What is the timeout?
• Is the result proved for one hub or for a network?

Terms

Virtual channel

An off-chain channel between two users, funded through a hub, that the hub does not update on every payment.

Ledger channel

The on-chain channel between a user and the hub, from which the virtual channel's funds are drawn.

The problem the paper names

A routed payment makes every intermediate node update its channels for every transfer. That costs liveness and attention from the hub. A virtual channel should let the two ends pay each other until they disagree or they close.

What the design proposes

• Two users each have a ledger channel with a hub.
• They and the hub open a virtual channel whose funding sits in those ledger channels.
• Payments inside the virtual channel are off-chain between the two users. The hub is not in the loop for each one.

How the mechanism is specified

• If everyone is honest, the hub learns that the virtual channel exists and then does not touch each payment.
• If someone misbehaves, the contracts on the underlying chain are the dispute path. The paper needs a chain that can express those contracts.
• The formal result is for one intermediary. A network of hubs is left as the obvious next composition, not as a theorem in this version.

What this page does not treat as proven

• This is not the Lightning paper and it does not run on Bitcoin's script as written. The model assumes richer contracts.
• A hub that disappears still forces a dispute. Virtual does not mean unsupervised.
• The paper does not establish the security of any later commercial hub.

Why a venture studio still reads it

Ask who must be online for a single payment. In Lightning, the route. In this paper, the two ends, until something goes wrong. Then ask which chain the dispute contract actually lives on.