elective-stereophonic
elective-stereophonic
New fees and size limits in 1.7
Please login or register.

Login with username, password and session length
Advanced search  

News:

Latest Stable Nxt Client: Nxt 1.12.1 Upgrade before block 2870000 is mandatory!

Pages: 1 [2] 3 4 ... 6  All

Author Topic: New fees and size limits in 1.7  (Read 21397 times)

Jean-Luc

  • Core Dev
  • Hero Member
  • *****
  • Karma: +816/-81
  • Offline Offline
  • Posts: 1610
    • View Profile
Re: New fees and size limits in 1.7
« Reply #20 on: November 11, 2015, 07:58:51 pm »

It is about fees for messages. But the catch is, because we want to have a free allowance of 32 bytes per message, to be able to attach e.g. order or account number to a payment without paying extra fee, if the fee for each 32 additional bytes is higher than the base transaction fee, one could cheat and split a long message into multiple 32 byte pieces, fitting each in the free allowance limit. For example, if fee per 32 bytes is 4 NXT, instead of paying 5 NXT for a transaction with 64 bytes message, the user will split it into two transactions with 32 bytes messages each and pay only 2 NXT, and this is worse for the blockchain than fitting it into a single transaction. So the base transaction fee must also scale up if the fee per byte is higher.

We could remove the free allowance instead, but since prunable messages necessarily leave behind a 32 bytes permanent hash, charging the user for a permanent message of less than 32 bytes doesn't make sense as moving to prunable will not help for such short messages, it adds more overhead.
Logged
GPG key fingerprint: 263A 9EB0 29CF C77A 3D06  FD13 811D 6940 E1E4 240C
NXT-X4LF-9A4G-WN9Z-2R322

cc001

  • Hero Member
  • *****
  • Karma: +68/-4
  • Offline Offline
  • Posts: 829
    • View Profile
Re: New fees and size limits in 1.7
« Reply #21 on: November 11, 2015, 08:27:14 pm »

We had a fee of 1 NXT when the price was 10 times higher, the market already took care of decreasing the fee. Why should we want to set it even lower? If the price goes up significantly, we will do another poll and adjust the fee. For now such adjustments will have to be done manually, and as each fee change requires a hard fork, are not practical to do more often than every 6 months, which is about how often we do hard forks now.

I totally agree with that. My only question is: Is the system able to use in the future (if needed) not only natural numbers of NXT as fee, but also fractions of a NXT as fees?
Logged
cc001 Personal - NXT-8RXS-2SSK-RNF2-HSNL8
NxtReporting.com - The Nxt Asset Exchange Portfolio Manager with Profitability Tracking - Donations are greatly appreciated on NXT-5W4G-GAR6-JHJP-H8ZTW

Jean-Luc

  • Core Dev
  • Hero Member
  • *****
  • Karma: +816/-81
  • Offline Offline
  • Posts: 1610
    • View Profile
Re: New fees and size limits in 1.7
« Reply #22 on: November 11, 2015, 08:30:05 pm »

Yes, fractional fees are fully supported, everything is in NQT internally. Although keeping it whole numbers may be easier for clients, and for end users.
Logged
GPG key fingerprint: 263A 9EB0 29CF C77A 3D06  FD13 811D 6940 E1E4 240C
NXT-X4LF-9A4G-WN9Z-2R322

NxtSwe

  • Hero Member
  • *****
  • Karma: +124/-9
  • Offline Offline
  • Posts: 657
    • View Profile
Re: New fees and size limits in 1.7
« Reply #23 on: November 11, 2015, 11:02:41 pm »

Yes, fractional fees are fully supported, everything is in NQT internally. Although keeping it whole numbers may be easier for clients, and for end users.
Never heard of milliNXT's or megaNQT's?  ;D
Logged
Check out the NxtLib, the .NET Framework API for the Nxt platform.

Tosch110

  • Hero Member
  • *****
  • Karma: +211/-18
  • Offline Offline
  • Posts: 2365
    • View Profile
Re: New fees and size limits in 1.7
« Reply #24 on: November 11, 2015, 11:56:33 pm »

Thank you Jean-Luc for your explanations! Very helpful.

I hope that the fee stays as 1 NXT so maybe integration can be more seemless and old feeNQT = 100000000 would still be accepted? ;)

crimi

  • Hero Member
  • *****
  • Karma: +122/-11
  • Offline Offline
  • Posts: 863
    • View Profile
Re: New fees and size limits in 1.7
« Reply #25 on: November 12, 2015, 11:37:18 am »



dont care about fiat value.
Logged

NxtSwe

  • Hero Member
  • *****
  • Karma: +124/-9
  • Offline Offline
  • Posts: 657
    • View Profile
Re: New fees and size limits in 1.7
« Reply #26 on: November 12, 2015, 03:09:19 pm »



dont care about fiat value.

To be fair, spending 1 NXT consumes the same amount of harddrive space and network bits as sending 1 million nxt.
It cost the same for me to take the flight with 1 EUR in my pocket as 1 Million EUR in my pocket ... well, with the exception of all the questions I would get from the security officers that is. But that's another matter.
Logged
Check out the NxtLib, the .NET Framework API for the Nxt platform.

crimi

  • Hero Member
  • *****
  • Karma: +122/-11
  • Offline Offline
  • Posts: 863
    • View Profile
Re: New fees and size limits in 1.7
« Reply #27 on: November 12, 2015, 03:26:47 pm »



dont care about fiat value.

To be fair, spending 1 NXT consumes the same amount of harddrive space and network bits as sending 1 million nxt.
It cost the same for me to take the flight with 1 EUR in my pocket as 1 Million EUR in my pocket ... well, with the exception of all the questions I would get from the security officers that is. But that's another matter.

Trust me if you want to send 1 btc and you have to pay 1 btc fee, bitcoin would be a dead drop today. This is impossible to market.
Logged

bcdev

  • Hero Member
  • *****
  • Karma: +162/-22
  • Offline Offline
  • Posts: 666
    • View Profile
Re: New fees and size limits in 1.7
« Reply #28 on: November 12, 2015, 03:33:43 pm »

To be fair, spending 1 NXT consumes the same amount of harddrive space and network bits as sending 1 million nxt.
It cost the same for me to take the flight with 1 EUR in my pocket as 1 Million EUR in my pocket ... well, with the exception of all the questions I would get from the security officers that is. But that's another matter.
You can't fly with more than $10k in US. Police can arrest you if you attempt that.
This includes any foreign currency. I've heard a rumor of a Canadian man who had 999 USD worth of CAD with him using the exchange rate of the day before. Sadly, the CAD/USD rose that day. Needless to say, he was arrested.
Logged

NxtSwe

  • Hero Member
  • *****
  • Karma: +124/-9
  • Offline Offline
  • Posts: 657
    • View Profile
Re: New fees and size limits in 1.7
« Reply #29 on: November 12, 2015, 03:42:45 pm »



dont care about fiat value.

To be fair, spending 1 NXT consumes the same amount of harddrive space and network bits as sending 1 million nxt.
It cost the same for me to take the flight with 1 EUR in my pocket as 1 Million EUR in my pocket ... well, with the exception of all the questions I would get from the security officers that is. But that's another matter.

Trust me if you want to send 1 btc and you have to pay 1 btc fee, bitcoin would be a dead drop today. This is impossible to market.

Are you saying that a $330 fee is too much if you want to send $330?
If so, we are in agreement.
Logged
Check out the NxtLib, the .NET Framework API for the Nxt platform.

lurker10

  • Hero Member
  • *****
  • Karma: +168/-33
  • Offline Offline
  • Posts: 1334
    • View Profile
Re: New fees and size limits in 1.7
« Reply #30 on: November 12, 2015, 03:44:49 pm »

To be fair, spending 1 NXT consumes the same amount of harddrive space and network bits as sending 1 million nxt.
It cost the same for me to take the flight with 1 EUR in my pocket as 1 Million EUR in my pocket ... well, with the exception of all the questions I would get from the security officers that is. But that's another matter.
You can't fly with more than $10k in US. Police can arrest you if you attempt that.
This includes any foreign currency. I've heard a rumor of a Canadian man who had 999 USD worth of CAD with him using the exchange rate of the day before. Sadly, the CAD/USD rose that day. Needless to say, he was arrested.

Land of the Free.
Logged
Run a node - win a prize! "Lucky node" project jar: NXT-8F28-EDVE-LPPX-HY4E7

dude

  • Full Member
  • ***
  • Karma: +44/-5
  • Offline Offline
  • Posts: 207
    • View Profile
Re: New fees and size limits in 1.7
« Reply #31 on: November 12, 2015, 03:45:39 pm »

Trust me if you want to send 1 btc and you have to pay 1 btc fee, bitcoin would be a dead drop today. This is impossible to market.

I don't see how having the fee at 0.5 for example could help this. The answer is, don't market it like that (you can express fees in fiat) or not at all, market useful stuff.

edit: Also, if I could get 150 BTC for $1, I wouldn't care.
« Last Edit: November 12, 2015, 03:52:11 pm by dude »
Logged

Damelon

  • Hero Member
  • *****
  • Karma: +792/-54
  • Offline Offline
  • Posts: 2314
    • View Profile
    • Nxt Inside
Re: New fees and size limits in 1.7
« Reply #32 on: November 12, 2015, 04:02:31 pm »



dont care about fiat value.

To be fair, spending 1 NXT consumes the same amount of harddrive space and network bits as sending 1 million nxt.
It cost the same for me to take the flight with 1 EUR in my pocket as 1 Million EUR in my pocket ... well, with the exception of all the questions I would get from the security officers that is. But that's another matter.

Trust me if you want to send 1 btc and you have to pay 1 btc fee, bitcoin would be a dead drop today. This is impossible to market.

This example only works with numbers that are large.

Listen to CoinOutlet: https://www.youtube.com/watch?v=FN0gUIVYu2s
They actually chose NXT because they thought the TX fee is low. And today, they would be right.

We already know that once the fees would become prohibitive (probably around 5 ct, but don't pin me down on it), the base fee could and would be changed. NXT is well prepared for that.

At this moment though, the fee in fiat is not the problem.

The only thing that could be against it is that relatively the fee is high, as when you consider them within the ecosystem.

This is a delicate balance though, because as long as the platform is cheap in marketcap, it's rather easy to misuse it when we'd lower the fees at this point.

We have never really had this issue crop up when talking to anyone. Mostly they say "what? 1 ct per tx? That's reasonable". Try to sell it in terms of fiat, and you're golden. Try to sell it in terms of the internal market of the platform, then you're in trouble.
Logged
Member of the Nxt Foundation | Donations: NXT-D6K7-MLY6-98FM-FLL5T
Join Nxt Slack! https://nxtchat.herokuapp.com/
Founder of Blockchain Workspace | Personal Site & Blog

lurker10

  • Hero Member
  • *****
  • Karma: +168/-33
  • Offline Offline
  • Posts: 1334
    • View Profile
Re: New fees and size limits in 1.7
« Reply #33 on: November 12, 2015, 04:05:15 pm »

Expressing fees in fiat works best.

If you calculate electricity/hardware fees, I saw a figure it actually costs $30 to send one Bitcoin transaction. These expenses are now subsidized through mining of new coins but what do you get in a few years when Bitcoin block rewards go down? NXT doesn't have this subsidy, on the other hand it doesn't require expensive hardware or paying for electricity much. 1 NXT has worked as a reasonable fee while 1 NXT < $0.10, let it stay this way. When the price goes above $0.10, the fees may become a burden, not yet. You can't expect to be able to write to a global ledger for nothing, it's a very limited resource. Hopefully CfB's Iota can enable micro transactions for milli-fractions of a US cent :)
Logged
Run a node - win a prize! "Lucky node" project jar: NXT-8F28-EDVE-LPPX-HY4E7

Audo Kryptowitz

  • Jr. Member
  • **
  • Karma: +69/-2
  • Offline Offline
  • Posts: 94
    • View Profile
Re: New fees and size limits in 1.7
« Reply #34 on: November 12, 2015, 04:58:22 pm »


Larger fees encourage people to secure the network.

The fees won't just vanish, rather they will go to those who provide a very important service with their hard earned NXT.
Logged
An Open Source Self-organizing Ecosystem - www.digitalcatallaxy.com

Hachoir

  • Full Member
  • ***
  • Karma: +12/-13
  • Offline Offline
  • Posts: 113
    • View Profile
Re: New fees and size limits in 1.7
« Reply #35 on: November 12, 2015, 05:04:09 pm »

Yes, perhaps more people will forge when it's profitable.
Logged

Brangdon

  • Hero Member
  • *****
  • Karma: +229/-25
  • Offline Offline
  • Posts: 1389
  • Quality is addictive.
    • View Profile
Re: New fees and size limits in 1.7
« Reply #36 on: November 12, 2015, 09:13:27 pm »

Low fees are killing forging economy and so just dangerous
I don't think they are, because whatever the fees, the income from forging is insignificant at current usage levels. From what I've read in previous discussions, the current fees are high enough to discourage use in third world countries, and for certain types of transaction such as updating a bid; and this is also dangerous.

So we disagree. We can debate it some, but eventually shouldn't we resolve this disagreement with a vote? If you succeed in keeping the options you dislike off the ballot paper, then that pre-empts the vote result. If you are right, then allowing the lower fee options is harmless, because the forgers will all vote for higher fees. (And, assuming it will be vote-by-account-balance, the big forgers are the only votes that matter.)

The other argument against low fees is spam protection. I'm more sympathetic to this argument. I'd still like to see lower fees, but maybe coupled to some other anti-spam mechanism, such as coin age or PoW. That could be some future update.
Logged

petko

  • Full Member
  • ***
  • Karma: +24/-0
  • Offline Offline
  • Posts: 100
    • View Profile
    • My blog
Re: New fees and size limits in 1.7
« Reply #37 on: November 12, 2015, 10:01:16 pm »

I still don't get it why minimal fees need to be enforced as part of the consensus.

Blockchain bloat will be better handled by a more adequate limit on the block size. Limit in bytes, not number of transactions. Like in Bitcoin. (Plus the impact each transaction has on the DB size can be included in the evaluation of each transaction).
I don't mean there won't be fees, I mean they won't be part of the consensus. Each forger will decide what minimal fees to require - it's better than voting with polls. And fluctuations in NXT price won't be a problem any more.

All unique names can be substituted by a nullable reference to an alias. We can have a currency or asset without name, and identified only by ID, why not? If the creator wants recognizable name, he/she attaches an alias. Aliases will expire. This way aliases belonging to lost account are not a problem any more, and there is no need of 25000 NXT minimal fee for a currency code.
Logged

Jean-Luc

  • Core Dev
  • Hero Member
  • *****
  • Karma: +816/-81
  • Offline Offline
  • Posts: 1610
    • View Profile
Re: New fees and size limits in 1.7
« Reply #38 on: November 13, 2015, 06:41:51 am »

I still don't get it why minimal fees need to be enforced as part of the consensus.

Blockchain bloat will be better handled by a more adequate limit on the block size. Limit in bytes, not number of transactions. Like in Bitcoin. (Plus the impact each transaction has on the DB size can be included in the evaluation of each transaction).
I don't mean there won't be fees, I mean they won't be part of the consensus. Each forger will decide what minimal fees to require - it's better than voting with polls. And fluctuations in NXT price won't be a problem any more.
We have discussed this before already. At the current transaction load, block size limit is not a factor that can influence the fees, unless we reduce it more than 10 times (and we do have two separate limits, both on transaction count and total transaction bytes size). Only when we get enough transaction traffic so that a significant percentage of the blocks are full, then a market rate fee can develop, by transaction senders willing to pay more in order to get their transaction included faster. With the current situation, if forgers can decide the fee it will gravitate towards zero, and expose the blockchain to spam. And we must think about the future, every MB of permanent data stored in the blockchain now is here to stay forever (the prunable side chains planned for 2.0 are not yet a reality and not guaranteed to be possible to implement). Migration to prunable data must be forced, leaving the permanent option only for those who really need it and are willing to pay the high fees for it.

Quote
All unique names can be substituted by a nullable reference to an alias. We can have a currency or asset without name, and identified only by ID, why not? If the creator wants recognizable name, he/she attaches an alias. Aliases will expire. This way aliases belonging to lost account are not a problem any more, and there is no need of 25000 NXT minimal fee for a currency code.
Could be. This is a theoretical discussion at this point, we have a working product and are not going to rewrite it in a different way that may or may not be better.
« Last Edit: November 13, 2015, 08:32:51 am by Jean-Luc »
Logged
GPG key fingerprint: 263A 9EB0 29CF C77A 3D06  FD13 811D 6940 E1E4 240C
NXT-X4LF-9A4G-WN9Z-2R322

Jean-Luc

  • Core Dev
  • Hero Member
  • *****
  • Karma: +816/-81
  • Offline Offline
  • Posts: 1610
    • View Profile
Re: New fees and size limits in 1.7
« Reply #39 on: November 13, 2015, 07:08:45 am »

Before someone complains that those fee and size changes have caught them unprepared and have not been announced in advance, here is a history of my posts on the topic, documenting that I have warned multiple times about this planned fees increase.

I first proposed that an increase of permanent AM fees should happen in 1.5.1e, when prunable data was implemented:

The size of regular plain and encrypted messages in this release has been
restricted back to 1000 bytes, and will likely be reduced even further, before
the stable release. This will be done in order to encourage users to switch to
prunable instead of regular messages. Fees for regular message attachments
will also be increased substantially.


We didn't do it for the 1.5 stable only because we decided to give people more time, until 1.6 (which was at the time expected to be a hard fork):

Quote
The size of regular plain and encrypted messages in this release has been
restricted back to 1000 bytes, and will likely be reduced even further, before
the stable release. This will be done in order to encourage users to switch to
prunable instead of regular messages. Fees for regular message attachments
will also be increased substantially.
It will break compatibility. I don't think that 1000 bytes limit should be lowered and the regular fee should be increased.
Of course it will break compatibility, but we can't stick with a wrong decision forever. 1.5 is expected to stay on testnet for long time. Even if we don't do the fee or size changes immediately in 1.5, they will have to be done in 1.6. One reason for reusing the same parameter names for prunable and permanent messages was to make it easy to switch, you only need to add messageIsPrunable=true.

1000 bytes for 1 NXT is not a wrong decision. You will have most of the messages prunable by simply making them default in the NXT client. And you can make prunnable messages fee slightly less to speed up transition to new messages. By breaking compatibility and raising fees you will only push away NXT services.
There are currently 130,000 transactions with unencrypted messages and only 10,000 with encrypted. The client defaults to encrypted, all plain ones must have come from API users. Changing the default to prunable in the client is not enough.

It is a tradeoff, the more we delay, the more API users will need to make the switch. Long term there is no excuse for keeping those verbose json messages permanently in the blockchain. I think permanent message size should eventually be limited to around 100 bytes for no extra fee, to fit things like order numbers, bitcoin addresses, account numbers or public keys. Above that, 1 NXT per 100 bytes with a hard limit of 1000.

The fees I gave were just an example and to start a discussion about it. But whether we increase fees now or in 1.6, or sometime in between, the long term direction has to be switching to prunable. Only data that the blockhain itself needs to verify and process future transactions should be permanently stored in it. Data that are only meaningful to external services, should be prunable after they have been spread around and recorded by archival nodes / service providers.

A few months later, the impending permanent message fee increase was mentioned again:

what is the size limit for "message" attachment that wont get pruned?
The limit for non-prunable messages has not changed, it is still 1000 bytes. Messages are pruned not based on size, but only if they have been created as prunable, i.e. messageIsPrunable=true parameter has been added. For prunable, the limit is 42 kbytes.

For 1.6 we should either reduce the size limit for non-prunable, or make them more expensive, to encourage the use of prunable messages instead.

And again:

You should run a few nodes with pruning disabled, where those are kept longer or indefinitely, from which older messages can be retrieved by those interested in them.

Expect the fee for permanent messages to increase significantly in 1.6, to motivate users to switch to prunable.


And again:

For 1.6, the fees for permanent messages will actually increase a lot. This is to force migration to prunable, there is no reason for people to continue using permanent messages now that prunable are available.

And again:

Concatenated arbitrary messages full of verbose JSON are another example of the fee not being a consideration. If it were, we would be seeing messages in binary, compressed, or at least minified json. Yet there is plenty of very verbose JSON in those, taking multiple messages instead of one, because it is more convenient for the programmers, and the fee is too low to make them worry about it. And they don't care about blockchain size either, or they would have tried to reduce message size even when it takes a single message only. To make them care, we will increase the fees for permanent messages.
Logged
GPG key fingerprint: 263A 9EB0 29CF C77A 3D06  FD13 811D 6940 E1E4 240C
NXT-X4LF-9A4G-WN9Z-2R322
Pages: 1 [2] 3 4 ... 6  All
 

elective-stereophonic
elective-stereophonic
assembly
assembly