👩‍🔧Great paper by @ClaraShik and @serg_tikhomirov on lightning jamming mitigations at eprint.iacr.org/2022/1454.pdf
🚨I expect all #lightning payments in future to have some tiny anti-spam fee attached (even if they fail): the paper suggests 2% of success fees (~0.001% of amount RN). That's looking at my node's mean fee for my LN payments of 0.05% (shockingly low, TBH, so I expect that to increase).

I told people to anticipate this at Lightning Conf in 2018, but that was an age ago, so let's revisit: why add a pre-pay fee?

If #Bitcoin confuses you, it's important to remember that we're trying to change *accepted social norms*. That's weird, slow and confusing.

In my Linux days we were dealing with "that's not how you construct software", Bitcoin is "that's not how you construct money". 1/7 For Linux I had read and sympathized with RMS's GNU Free Software vision at University, but it took Linux to actually show me it actually *worked better*. But I was the "weird Linux guy" for a long time, then inside a "weird Linux bubble". 2/7

So @alexbosworth doesn't understand the Lightning Spec process. I thought Lightning Labs had been so quiet in the last two years because (1) @roasbeef is the only one there who's allowed to do spec stuff (2) profitability lead them to pivot to proprietary centralized offerings. Turns out they've decided they can leverage network dominance into protocol control, and the spec process isn't "real".

Mea culpa: I have spent far too much time on the spec process, and not enough on implementation polishing and marketing. That changes today.

There are three types of spam problem in #Bitcoin #LightningNetwork; let's call them "local", "slow" and "fast". local spam is someone sending or requesting too much data: that's a Simple Matter of Code, with ratelimiting etc. Let's talk about the other two: "fast" spam is sending many failed payments. You can't tell what will succeed, and you can't blame the sending peer (they may not be the source). This attack is free since we only pay on success. The answer is to require an up-front payment, but

I used #clightning's new multifundchannel with every one of the 1837 nodes on testnet. Most are dead, many timed out, some already had channels, and two didn't respond to open_channel, but 106 made it! Here's the open tx:
blockstream.info/testnet/tx/cde… Thanks to @niftynei for the freshly mined testnet coins, @ZmnSCPxj for the new multifundchannel plugin. We've been able to do this in theory for over a year, but now it's trivial! Though I had to use python: bash told me "Argument list too long".

lnprototest is vital for lightning development. It's a simple place to rapidly implement new features for testing: very hackable! It also lets us test things that real implementations don't do. It's an easy on-ramp for interoperability testing.

github.com/rustyrussell/l… For example, I accidentally broke #clightning accepting unknown odd packets. Nothing broke! But this is vital for future upgrades, where we want to rely on this "you will ignore this, not barf" behaviour. lnprototest's predecessor found this.

Congratulations to my father, Dr John Russell, for receiving the Order of Australia (AM). "For significant service to medicine in the field of anaesthesia, and to medical health standards.". 1/4 He set standards, in all ways. He literally wrote the book on Anesthetic Equipment in Australia, he has a deep sympathy for human errors, and spent a lifetime's work understanding and limiting their causes. He was the anaesthetists' anaesthetist. 2/4

As a happy @TheRealBlockFi user, I highly recommend you join me in withdrawing all your funds over Jan 3 for #ProofOfKeys. On additional to the normal "check they are solvent" and "firedrills are good" reasons: 1. Anyone can claim to pay interest: a-la OneCoin, the proof is in being able to withdraw it. You need to verify this occasionally.

Due to a problem with Ubuntu not really caring about bitcoind's PPAs, you're supposed to use the SNAP now. That's an undocumented disaster. 1/ 1. /usr/bin/bitcoind -> /snap/bin/bitcoin-core.daemon, similarly bitcoin-cli.
2. No you can't create a symlink, because they use the name to distinguish.
3. Hey, the default directory silently changed: it's now /snap/bitcoin-core/common/.bitcoin

ICYMI: Prelude to my #TLC2019 talk was two cautions: 1. Fees will rise on LN. 2. Long-held HTLCs will be penalized. Let me tweetstorm that at you now... LND's default fees are 1 sat + 1ppm. The proportional part can only go down to 0; we're already seeing that raised by nodes, and there's clearly room for this to grow. 100ppm == 1 basis point, so we're 100x smaller than the fin community can measure!

#TLC2019 was over too fast! Everyone I spoke with was trying something exciting, but I feel I barely scratched the surface. Got to fanboy @JackMallers far too briefly (hey, you still in town?). Main things on my to-do list below: 1/ Offers. Well received, great feedback and ideas, main q was "when?". Going to take at least 6 months, but I'm going to start now. Watch mailing list for proposal.

Finally got around this; nice to read some non-trivial #bitcoinlightning analysis, am enjoying!

https://twitter.com/pedrorechez/status/875821488552718336

@pedrorechez @roasbeef 1/ What we refer to as "payment decorrelation" seems better handled by using EC points: see lists.linuxfoundation.org/pipermail/ligh…