What happened this month

https://arkdev.info/blog/bitcoin-virtual-channels/
One of teams developing Ark added a type of payment channel called a “Spillman Channel” to their implementation
I’ve decided to take this opportunity to tell you what a Spillman Channel is and how it differs from a Lightning Channel

Spillman Channels versus Lightning Channels

Equal: Spillman channels and lightning channels both let you avoid putting your transactions on the blockchain
Equal: As a result, both are usually a lot faster and cheaper than using the base layer
Benefit: Spillman channels are simpler; lightning channels are more complex
Benefit: In Spillman channels, wallets only have to store the latest state; in lightning channels, they have to store every state (at least parts of them)
Benefit: Spillman channels have no “inbound capacity” problem, resulting in fewer payment failures; lightning channels do have that problem
Drawback: Spillman channels are unidirectional (“send only”); lightning channels are bidirectional (“send and receive”)
Drawback: Spillman channels must be closed after a fixed number of transactions; lightning channels can stay open as long as the users want
Miscellaneous: Spillman channels were conceived and implemented in 2013; lightning channels were conceived in 2015 and implemented on bitcoin in 2017

Like lightning, two users create a 2 of 2 multisig address
Like lightning, one user deposits money into that address
Before doing so, the depositor gets a transaction signed by his or her counterparty letting the depositor withdraw 100% of their money after a 30 day timelock
Like lightning, all withdrawals from the multisig involve two transactions: the first starts the countdown, then, when the countdown is over, the user does the “actual” withdrawal in a second transaction
The first time the depositor wants to send money to the other user, they sign a new transaction sending part of their balance to their counterparty, minus 1 day (so the timelock is for 29 days)
The second time, they do it again, giving even more money, but minus 2 days (so the timelock is for 28 days)
The depositor can keep doing that til all 30 days are used up, then they must close the channel. They may make a new one if they want to keep it up

The depositor can never broadcast an “old state” in a Spillman channel without giving their counterparty an opportunity to broadcast the “latest state” (because the “latest state” becomes valid first, thanks to its shorter timelock)
The depositor also never has to worry about the “other user” broadcasting an old state, because that would just give the depositor some money back – remember, every “new state” sends more money from the depositor to the other user, since Spillman channels are “send only”

The team that added Spillman channels wants their version of Ark to be a “consumer app” – and consumers receive bitcoin infrequently
Ark has a built in method for sending and receiving transactions cheaply, but they take 10 minutes to reach finality
Spillman channels give “senders” instant finality without introducing LN’s drawbacks (complex state management and inbound capacity constraints)
Since they expect their users to “receive” infrequently, they think it’s okay for that part to still take about 10 minutes to reach finality