Your daily (lunar) base where you will find all the lightning-related news outer space.
Lightning is a Moving Target!
This week, in Episode 33, we talk to Kevin Cai, an infrastructure and system engineer working at Kollider. He ensures that all systems at Kollider and their own Lightning Node running continue to run smoothly at all times!
Kollider is a real-time settled derivative exchange. Their trading and settlement system uses the Lightning Network to enable traders to get instant price exposure without having to pre-fund a trading account.
Among a lot of things we talk with Kevin about all the things Kollider is working on, like gamification with trading competitions, where you can win sats as a prize every month.
He also shares his thoughts about running a stable Lightning Node and about the future of the Lightning Network!
Watch or listen to the full interview in this brand new show or start reading below!
Enjoy, but beware of Kollider’s addictiveness!
Do you like this article and our other content? Donate and support us in our mission to create a worldwide distributed Lightning Network by connecting countries around the world and bringing The Lightning News to you first!
Follow The Daily Moon also on Telegram and Twitter!
What was your first experience with the lightning network and do you still know what you did/saw at that moment?
My friend Jacob showed me an app called Breez and demonstrated sending $5 to me. This was after he had explained the basic concepts of the LN so I knew what was going on. I was quite impressed.
A year ago you said that Lightning is in your opinion the single best UX improvement to bitcoin that exists and that Lightning is most intuitive. Can you emphasize again why you think this?
Bitcoin on-chain is a bit of a nerd’s game. You need to know what a UTXO is, what inputs and outputs are for a tx, and generally understand that until that TX is confirmed, it’s not really there. This can create huge issues for developers that are used to fiat
– In my opinion, lightning as an abstraction layer makes adoption easy on the developer, and usage easy for the user.
Do you think Lightning is more difficult to grasp for newbies than Bitcoiners who fully understand Layer 1?
In the beginning, yes. But I think additional educational resources should simplify this. The issue I think is that lightning is a moving target, so you can’t really get your degree on it so to speak. You need to keep up with its feature updates and such.
– Example: AMP invoices, bolt12 offers, etc.
Do you think that the Lightning Network needs possibilities for Yield generation to grow more adoption?
Perhaps. I haven’t given it too much thought. For me, the idea of selling channels is more or less a form of yield generation.
Can you pay with lightning where you live? Are there merchants that accept BTC or lightning?
What features does lightning need to reach the next level of maturity?
– offline payments
– bulletproof backups
What is on the top of your list of Bitcoin / Lightning concepts to study next?
– PTLCs (conditional payments that can replace the use of HTLCs in LN payment channels)
– By contrast, HTLCs: are locked using a hash digest and unlocked by providing the corresponding preimage.
– The most commonly used hash function is SHA256, which produces a 256-bit (32-byte) digest commonly generated from a 32-byte preimage.
– all payments use the same preimage and hash lock. This creates a link between those payments if they’re published onchain or if they’re routed offchain though surveillance nodes.
– PTLC point locks: are locked using a public key (a point on Bitcoin’s elliptic curve) and unlocked by providing a corresponding signature from a satisfied signature adaptor. For a proposed schnorr signature construction, the key would be 32 bytes and the signature 64 bytes.
– However, using either multiparty ECDSA or schnorr key aggregation and signing, the keys and signature can be combined with other keys and signatures needed to authorize any spend, allowing point locks to use zero bytes of distinct block space.
Not so long ago you wrote that one factor to consider for dynamic fee setting programs is that you could trip LND’s heuristic for a peer that updates their fee too often. So it’s not always just about the fee level, but also how often it changes. Do you use a program for dynamic fee settings? And can you explain this trade off you were talking about?
Yeah so I spoke to Alex Bosworth a few months ago and asked him about dynamic fees. His feedback was that lnd uses a set of heuristics to determine how much a peer should be favored over another arbitrary peer for selection in routing.
– If you have a low fee, then you route a payment and raise the fee (so if your local side went down) that could make the other party’s lnd mark that channel as bad. Of course, this doesn’t mean you can’t update fees. Just that you shouldn’t do it all the time, like once every five minutes.
– Personally I use charge-lnd running on an hourly basis
Smallworlnd said and I quote: “My approach to it is that I’m trying to set fees to reflect the value of liquidity availability. It’s mostly one-size-fits-all because I dislike mis-pricing liquidity more than the potential loss of routes because the fee is too high.” He prefers to maintain consistent fees and refrain from gossiping too much.
In this early phase of the Lightning Network, isn’t it very difficult to properly estimate the availability of liquidity? How will this change when there are more merchants responsible for an increase of transactions?
Yeah I agree with him there, I don’t think you should be setting misleading fees.
– It’s actually not too hard to estimate liquidity if you build tools to do probing. Sidenote, this is how tools like loop work.
You experimented or are still so with adding a data source in the form of probing channels for attempted payments. This is just one of the metrics to help you to make better decisions.
Probing is used a lot within the Lightning Network for several purposes. If a lot more Nodes operators are probing to get more data they need: Do you think this can be a problem for the speed and stability of the
This should be less of a problem over time; right now, lnd has a nasty habit of using lots of disk. Even for failed payments. So as your database just grows monotonically, naturally this can create operational problems.
– Eventually this should be fixed. I think lnd 0.15 has some big changes in this aspect.
Which programs do you use on your node to automate for example fees or rebalancing? And which ones do you recommend?
– bos (for telegram notifs)
You have your node running on clearnet. What is your setup to achieve this in a safe way and why did you choose to not run behind tor?
Great question. I run my node clearnet behind a proxy, so the IP advertised via gossip protocol is not my home IP. My services run behind cloudflare to obfuscate the IP as well.
– I actually also listen behind tor as well, to provide hybrid connectivity.
You calculate good peers to open a channel to, making use of LnNodeinsight from Smallworlnd (Episode 11). What do you use besides that to search for the best next Nodes?
I have this other tool by LnNodeInsight which simulates and min-max optimizes channel opens for nodes that my node doesn’t have a channel to already. It calculates the change in centrality.
When starting a Lightning Node there are a lot of software distribution possibilities to choose from. For example Rapiblitz, Umbrel, Citadel, Road to Node and a lot more. Your choice often also depends on your level of knowledge and experience. What do you think is best suitable when starting as a Node Operator?
Kind of a controversial opinion here, but I really like Umbrel despite its license. I’ve talked to Luke before, he’s a really knowledgeable and nice guy.
Some people have big nodes but not much knowledge of the command line. When is it absolutely necessary to have more experience with the command line? Does this depend on the strategy someone has with his Node?
I would say it definitely doesn’t hurt to have some CLI experience. Sometimes there isn’t a GUI available, and you have to make do. Sometimes, the GUI shows less info than the CLI does. It really depends on your use case but in general I think there’s value
in spinning up a Linux VM and playing around with it. I learned my Linux fu with minecraft server
If the Lighting Network will eventually scale to a new infrastructural standard, do you think centralization is inevitable?
To a certain degree, absolutely. When you look at the graph nowadays, the thing that you tend to notice is that when an exchange joins the LN, it makes big waves. That might be because they have access to lots of large UTXOs, and it might be because they can
stick a person to be responsible for running and managing that node. Either way, there are forces at work that make it easier to go from zero to hero.
Is centralization always a bad thing in your opinion?
No, I don’t think centralization is always bad. It depends on the intent. If there is a group, say like the NSA or another nation-state 3 letter agency trying to analyze traffic, that’s bad. If a bunch of exchanges want to get into lightning and make it better
but also centralized, that’s not bad. There’s no reason why independent whales can’t come in and undercut the misbehaving exchanges (in theory). In my opinion it’s just growing pains. Ultimately decentralization should result from capital being deployed into LN.
How do you envision the future of running or having a Lightning Node of your own?
Paid seeding on torrents — I feel like this is a missed opportunity (talk about Web3Torrent)
Around the BIP119 Miner Activated Soft Fork, proposed by Jeremy Rubin, there was a lot of discussion. CTV allows for the implementation of covenants. Covenants are useful for constructing smart contracts with several use cases, such as preventing your funds from being stolen in case of hacking and helping scale the network.
What were your thoughts about this proposal?
So I really like the idea behind BIP119, but I think the activation of it has become quite polarizing.
– Some of it is due to misunderstanding how it works (like saying that it would allow govts to create addr whitelists), and others just think the change is being rushed. I’m not super opinionated on it tbh.
In the last episode we talked with Dr. Olga Ukolova of Pandora Core. She is developing 2nd layer solutions for scalable smart contracts with RGB. The most interesting thing was we have to see RGB as something completely different from Bitcoin or Lightning.
Did you read into what RGB is and what do you think of the possibilities of RGB when thinking of the future and developments of the Lightning Network?
Yes. I find the idea of transferring fungible assets over lightning to be very exciting. I think it could bring some much needed attention from people who would otherwise only be interested in ETH to the BTC ecosystem.
– Another controversial opinion incoming. I think Taro will probably outpace RGB because of its funding and association with LL
What’s the most bitcoin/lightning friendly country in the world (except from El Salvador) ?
– There are no special laws or regulations for cryptocurrencies in Japan
– as long as they abide by anti-money laundering (AML) requirements and other existing financial rules
We have LND (lightning labs), c-lightning (blockstream), eclair (Acinq) and some more… These are all implementations on the second layer. What lightning implementation do you prefer/use at the moment?
– such as those set out by the Financial Services Agency (FSA).
I use LND, but I have tried core lightning and eclair. They have their own strengths and weaknesses for sure.
– For eclair you can do fiat hosted channels, which I actually use all the time. Shout out to StandardSats.
If you look back to the last 2 years of lightning development and have to look 2 years in the future where do you think we are with lightning by then?
Not sure tbh, lots of possible ways we can make lightning better