I don't think the same thing happened to SolidCoin 1.0. SolidCoin 1.0 was using a fixed fee system, and ArtForz took advantage of that and spammed the network and only paying the same low fixed fee as everyone else. So RealSolid has to shutdown the chain because his change to Bitcoin's fee system backfired. Having a fixed fee is good, but that's assuming no one abuses it and sends spam transactions for low cost.

Let me summarize what's been happening

Dust spamming transactions

Litecoin is using the SAME Bitcoin code to deal with spam. The problem was that when I initially released Litecoin, I tried to change as a little as I could from Bitcoin. So I naively left the fee the same: 0.0005 ltc for transaction fee and 0.0001 ltc for transaction relay fee. This made it too cheap for a spammer to bloat our chain. The spammer started off by sending transactions with a ton of 0.00000001 LTC outputs and paying very little fees. (See: http://blockexplorer.sytes.net/block/4d5fc716d8a56f4223d38ebdaa2858551302d2118084448895db4131a3087117) So I increased it 200 times to make it closer to Bitcoin dollar-wise: 0.1 ltc for transaction fee and 0.02 ltc for transaction relay fee. See these 2 commits:

https://github.com/coblee/litecoin/commit/b661930c59691761691e0690771981af37b2b835
https://github.com/coblee/litecoin/commit/f0f9b11e860acb96279f26010c55d30b0684f10b

Exempt low-priority transactions

After this change went in, the spammer decided to send a lot of 2 LTC transactions for free. (See: http://blockexplorer.sytes.net/block/cf33787eec7468824617350664ec77048e6ecf47e6ade82a3bf07d5d566219a7) Bitcoin allows 3000 bytes of low priority transactions to go into a block for free. So he manages to fit about 13 of these transactions into each block without paying any fee. It's annoying but not too bad. But since we are finding blocks 4 times as fast as Bitcoin, this attack is 4 times as effective than a similar attack on Bitcoin. So I made the change to reduce the amount of exempt free transaction space to 500 bytes:

https://github.com/coblee/litecoin/commit/65f2053c72522a12ff92429c28d654ac8002cea4
https://github.com/coblee/litecoin/commit/332adc52190d079ccc3a6da0457ad2a494fc3866

I also changed the priority threshold to match Litecoins parameters. It was previously using 144 for the numbers of blocks found a day. So by using 576, the priority cutoff is now 1 litecoin day per 250 bytes.

https://github.com/coblee/litecoin/commit/8fe8e1304c7e7ed15e5499054e24b97b96ba98a1

High-priority transactions

Once that fix went in, the spammer can no longer fit 13 of those low priority transactions in each block. He can fit about 3. So he decided to make his transactions high priority. He started sending 577 coins to himself per transaction. (See: http://blockexplorer.sytes.net/block/960d4b540de8663508fad8e3a24b8329bc8c435a1233648281788a557ccbdcfb) He could then fit about 13 of these into a block. Since they are high priority, they can be sent without any fees. And they are only limited by the free transaction rate limiting code to prevent abuse. I then decided that we can live with this since honest user's transactions will be higher priority and will go in ahead of this spam. This attack will just slowly bloat the chain.

Dust spamming transactions round 2

I think after the spammer saw that I wasn't going to do anything about the high-priority transactions, he decided that's it's not worth his time to continue that since it's not doing much damage. He went back to dust spamming. But this time, he's taking advantage of the lower relay transaction fees and the fact that many users have not upgraded to the latest code. So he's paying 1/5 the cost that he normally should be charged. He's paying about 1 LTC for 50k of chain bloat. (See: http://blockexplorer.sytes.net/block/fa27fbea0d3e1a45021debb64c894ad619223baa5ee594c6fd47cda4d3e5168c) This is the problem we currently have.

This exact same attack can be performed on Bitcoin, but I don't think anyone is wasting time with it. You can bloat the Bitcoin chain by 50kb by spending 0.025 btc or 6 cents. (and 1GB is only going to cost $1200) My fix proposed above should fix this problem for both Litecoin and Bitcoin.

Here's another thought I have. Ideally, I'd like to tie the transaction fees with current price of Litecoin so that the cost of spamming stays the same in terms of fiat cost. (We are still living in a fiat world whether we like it or not.) But that's impossible to do in code. The next best thing is to tie it to the difficulty. The idea is that difficulty is loosely coupled with price. If ltc price goes up, difficulty will follow as more miners jump in. And as ltc price goes down, mining becomes unprofitable, miners leave, and difficulty goes down.

So we want an inverse relationship between the base fee and the difficulty. When Litecoin becomes popular and more people start to mine it and ltc price goes up, we want the fees to go down. And we should also cap the max fee at 1 LTC. Something like this.

BaseFee = 0.1 difficulty (round to 2 significant digits)

With the current difficulty at 0.63769, the base fee will be 0.15 LTC.
If the network doubles and difficulty doubles, the base fee will be reduced to 0.075 LTC.
And if we ever go back down to difficulty of less than 0.1, the base fee will be at most 1 LTC.

Thoughts?

I'm not sure difficulty is that much linked to coin value. For instance, the BTC diff is now about what it was in early June or something, but its value is like 10 times lower.
I'm not sure either if a fluctuating fee (or to put it in other words, a somewhat unpredictable fee) is a good thing psychologically. Manual adjustment from time to time, depending on the economic situation, sounds good enough to me.

It's not really centralized since noone is forced to update their client.
If I understood well, the point of coblee's idea is precisely to keep a stable fee (in fiat value), which is somewhat equivalent to the fees going down as the network becomes more secure.

Note that I don't really mind the automatically adjusting fee, I'm just not sure it's as needed as psychologically uncomfortable. When I started using BTC, that mysterious unpredictable fee was a major turn off to me.

Wow, I go on vacation for a week and come back to see this mess!  On the one hand, fuck RS or CH or whoever the fuck it is that's doing this, but I've also been very happy to see the innovative fast response and reasonable client changes as a result.  Thanks for all the hard work coblee, litecoins rock!

It's really hard to figure out your position on stuff, bulanula. And plus you seem to flip flop so often. Ever thought of running for a political position?

Are you against manual transaction fee adjustments or for them? Do you know that the bitcoin fee has been manually adjusted from 0.01 btc to 0.0005 btc as the bitcoin price rose to $30? And now that the price is back to $2, the current bitcoin fees make it very cheap to spam bitcoin also.


Yes, so if bitcoin had auto adjusted fees, the fee won't increase back to 0.01 btc because the difficulty didn't drop so much. But it would be about twice the 0.0005 btc fee, which would at least be going in the right direction without having to release another client just to manually adjust the fees.


Yes psychologically, it might be bad. Though it's just a simple calculation that anyone can do if they know the difficulty, so it's not something mysterious.

Well you can just use allchains to get an estimate of next difficulty. Or if predictability is that important, I can always fix the fee to the previous period's difficulty. So you will just use today's difficulty to figure out the fee for the next period. Like everything it's a tradeoff. Would you rather have to upgrade every now and then to get new adjusted fees? Or would you rather have fees set automatically to something close but then it's not something that's fixed.

Sounds good indeed. And maybe to reduce variability (improve predictability even further), if it's not too much trouble, compute today's fee based on average diff of the last 2 weeks or something (like average of the last 4 diff).

Why not cap fees where they are right now and use formula to decrease them as difficulty rises.  With current fee structure spamming would only make sense if value of LTC fell 90% or so.  If LTC value has fallen that much the risk of fees being too low are likely not really a significant concern.

It may be easier to convince people fees can only fall than fees can rise and fall but it is unlikely they will rise.