[NXT] NXTInfrastructure committee - page 2.

mczarnek

hero member

Activity: 527

Merit: 503

Quote from: marcus03 on March 11, 2014, 03:38:22 AM

Quote from: EvilDave on March 10, 2014, 07:18:14 PM

I'm with Marcus here, lets keep most of our biz here on the forum, in the public eye.

Moving on to business:

Secondleo sent me a copy of a paper on energy consumtion and efficiency compared between the BTC and NXT networks.

https://docs.google.com/document/d/1J8uhdshu9epGRrQHBaloGc4itdvuAHZDAUtNDjOhz-8/edit?usp=sharing

I've had a quick look thru it, and it is full of useful information, might deserve some bounty even as a stand-alone project.
But SecondLeo also has a further proposition/project to do with an efficient forging soution.
He got in contact with me via Nextcoin.org:

I am one of the authors of the recently published paper about the energy efficiency of the Nxt network.

On the basis of the work put into the paper I have plans in motion to offer a powerful and minimum effort solution for forging.
I would appreciate it if you provided a point of contact for the committee so I can present the project there and see if I can get any support.

i'd like to hear more, so I've invited secondleo to come on over and tell us everything....

Generally, I think the paper might deserve funding.

Would you please set up an Issue for this in the issue tracker at: https://bitbucket.org/nxtinfrastructure/committee/issues/new
(Set the "component" to "Projects/Bounties to fund".)

Hey, I'm the other author on that paper, we're still trying to figure out how to reasonable estimate how many forgers we'll have forging for Nxt, if anyone has any thoughts I'd love to hear them. I'm thinking maybe we should incorporate 5 servers that have cloudflare protection for DDOS attack? I understand that's our current solution?

Regarding which machines to recommend people forge with, that is SecondLeo's area.

Also, you should know, Salascz did give us 5000 Nxt as a bounty for the paper already.

I'd love to discuss it with anyone who would like to. We're still tweaking that one last number then are probably ready to publish it and hope to spread around, say publish it on coindesk and other places that would allow it to market Nxt, up the price, and hopefully make these committee funds more valuable.

EvilDave

hero member

Activity: 854

Merit: 1001

Just opened threads on Nextcoin and Nxtcrypto for the InfCom:

https://forums.nxtcrypto.org/viewtopic.php?f=2&t=881&p=4429#p4429

https://nextcoin.org/index.php/topic,4449.0.html

EvilDave

hero member

Activity: 854

Merit: 1001

Infrastructure Committee decisions.

Girls/guys/undecided:

The InfCom has been seriously busy over the last few days, and we have finalised two of the issues presented to us.

Most of our decision-making process can be seen on BitBucket, with some debate/input happening here on BTT.
PMs have been kept to a minimum to ensure complete transparency, and any relevant info from PM exchanges has been posted either here or on Bitbucket:
https://bitbucket.org/nxtinfrastructure/committee/issues?status=new&status=open

On to the decisions:

After a lot of debate, InfCom has rejected opticalcarriers request for funding for SSL certification for VPS, we feel that TOR provides a better level of security/privacy:
https://bitbucket.org/nxtinfrastructure/committee/issue/30/public-api-nodes-could-support-ssl-for

After even more debate, InfCom has approved an initial bounty payment of 2000 NXT to secondleo (+Matthew C, collaborator on the paper) for his paper on energy efficiency in the NXTwork:
https://bitbucket.org/nxtinfrastructure/committee/issue/19/nxt-energy-efficiency-paper-secondleo
There is a further 3000 NXT bounty ready to go for an improved version of the paper, based on feedback from InfCom.

That's about it for today, keep up the good work. Any questions, feedback, requests, get in touch.....

NXT Infrastructure Committee members:

EvilDave (spokesthing)
Marcus03
chanc3r
ChuckOne
ferment [/i]

marcus03

full member

Activity: 224

Merit: 100

Quote from: EvilDave on March 18, 2014, 01:34:00 PM

My first question, is nxtcrypto.org/SSL/wiki an InfCom issue ?

I'd say yes.

Regarding SSL for the nxtcrypto.org wiki, no, this was not discussed when we discussed opticalcarriers application, so we should probably do it now.

Generally, the problem with exit point operators being able to read the unencrypted communication going through them in the Tor network is valid.

I don't understand though what exactly needs to be protected when editing a public Wiki. I don't see the security problem (and there definetly is no privacy problem in regards to tracking the editor's IP when editing the Wiki through Tor).

EvilDave

hero member

Activity: 854

Merit: 1001

The follow up SSL discussion, this time as to whether we need SSL on nxtcrypto,org to provide extra security for the wiki.

The text below was originally from PM:

Quote from: marcus03 on March 18, 2014, 03:01:55 PM

Hi!

Can we move this to the InfCom thread at https://bitcointalksearch.org/topic/nxt-nxtinfrastructure-committee-506757 instead of a random list of pm recipients?

Cheers,
Marcus

Quote from: EvilDave on March 18, 2014, 02:53:07 PM

Ok, hold on a cotton-picking moment, guys.

We came to the conclusion that SSL wasn't necassary on the VPS, that TOR was a much better option.

I've had a look thru the Bitbucket again, and I cant find anything about SSL and wiki-editing on nxtcrypto-org, apart from the title of this issue, which then goes on to be all about VPS:
https://bitbucket.org/nxtinfrastructure/committee/issue/20/ssl-certificate-for-nxtcryptoorg
Has this issue been discussed elsewhere?

The wiki-editing/SSL on nextcrypto seems like a valid concern, and is a seperate issue to the use of SSL on VPS.
I'm going to cc this PM to the rest of InfCom, see if we should open another issue on this and consider funding from the point of view of wiki protection.

So, don't give up yet, OC.

EvilDave.

Quote from: opticalcarrier on March 18, 2014, 01:46:19 PM

Joefox,

They rejected my request for SSL certs for the domain. They are recommending use of TOR for the VPSs, which isnt a bad idea at all (I use tor for IP privacy) but for people who use TOR (like me)and log into the wiki for editing, SSL becomes a CRITICAL issue since the exit tor node sees all the traffic UNENCRYPTED. So tor users who are also wiki editors when the wiki doesnt have good SSL leave a gaping security hole that we really need to plug.

W/O a real CA-signed cert the best we can do is, on the wiki, load the wildcard cert I created and signed with my own private self-signed cert. But this will confuse the wiki editors as it will give security errors out to all wiki editors telling them that someone is doing something fishy with the connection, unless they do the extra step of loading my private CA cert into their browser. Get with me an I can provide both the wildcard cert for your wiki server as well as the private CA that we need to distribute out to all wiki editors.

-OC

Quote from: EvilDave on March 18, 2014, 01:24:38 AM

Optical:
After a lot of debate, InfCom has come to the conclusion that although SSL would help with the perception of security, it wouldn't add that much more actual security/privacy.

The concensus seems to be that we should concentrate on using NXT over TOR for added security.

So, in my role as InfCom spokesbeast: thats a "no" to your SSL funding request. Sorry.

U can see the decision-making process here:
https://bitbucket.org/nxtinfrastructure/committee/issue/30/public-api-nodes-could-support-ssl-for
Feel completely free to object if you like.

Good luck, keep up the good work,

EvilDave.

My first question, is nxtcrypto.org/SSL/wiki an InfCom issue ?

joefox

full member

Activity: 210

Merit: 100

Quote from: EvilDave on March 17, 2014, 03:32:18 PM

Quote from: NxtMinnow on March 17, 2014, 11:30:20 AM

Good morning Infrastructure Committee,
My Proposed Goals:
Nxt AE release must be coordinated with an effective Marketing Campaign.
Nxt needs to (IMMEDIATELY) put together proper advertisements for programmer jobs complete with description of required duties and offered pay.

Just my feedback here:

AE/marketing: totally +10 with u here, but this is TechDev/Marketing Commitee territory, not InfCom.

I have no issues with NxtMinnow's proposed goals -- what it means is that we have to coordinate between committees, that's all. IMHO marketing is really a "downstream" activity -- we promote things that DO exist, not things we think SHOULD exist. The latter gets us into trouble.

EvilDave

hero member

Activity: 854

Merit: 1001

Quote from: NxtMinnow on March 17, 2014, 11:30:20 AM

Good morning Infrastructure Committee,
This is a brainstorming session. We stand at the helm of the most efficient blockchain technology available. We need to set a course. We need a roadmap and timeline of Nxt goals.

My Proposed Goals:
Nxt AE release must be coordinated with an effective Marketing Campaign.
Nxt needs to (IMMEDIATELY) put together proper advertisements for programmer jobs complete with description of required duties and offered pay.
Nxt needs a USDollar to NXT registered exchange, preferably in New York State.
Nxt Point of Sale Use Case development.

We should open subthreads once the first top one or two priority projects on the roadmap are identified.

NXT Infrastructure Committee please brainstorm your ideas for Nxt infrastructure goals.

Just my feedback here:

AE/marketing: totally +10 with u here, but this is TechDev/Marketing Commitee territory, not InfCom.

Dev team needs to be beefed up. Again more a TechDev/whole community issue.

Dollar/euro/fiat gateways, yes. Not sure about basing them in the US tho', the Yanks might just change the rules in 2 weeks time and kill the whole deal.
I'd vote for a more financial freedom (and freedom in general) loving country such as Switzerland or Leichtenstein.
I believe Ethereum has gone the Swiss route.....

Point of sale: +100. Like RJ keeps on banging on at us, NXT should be able to succeed as a 1st gen crypto, if only to give us more money to support 2nd gen development.

Ans as marcus03 has just pointed out, the InfCom brains are being stormed.

NxtMinnow

member

Activity: 84

Merit: 10

Great work InfCom. Thank you for your commitment to maintaining excellent transparency in Nxt InfCom.

marcus03

full member

Activity: 224

Merit: 100

Quote from: NxtMinnow on March 17, 2014, 11:30:20 AM

Good morning Infrastructure Committee,
This is a brainstorming session. We stand at the helm of the most efficient blockchain technology available. We need to set a course. We need a roadmap and timeline of Nxt goals.

My Proposed Goals:
Nxt AE release must be coordinated with an effective Marketing Campaign.
Nxt needs to (IMMEDIATELY) put together proper advertisements for programmer jobs complete with description of required duties and offered pay.
Nxt needs a USDollar to NXT registered exchange, preferably in New York State.
Nxt Point of Sale Use Case development.

We should open subthreads once the first top one or two priority projects on the roadmap are identified.

NXT Infrastructure Committee please brainstorm your ideas for Nxt infrastructure goals.

Have you checked what InfCom has been done until now?

https://bitbucket.org/nxtinfrastructure/committee/issues

NxtMinnow

member

Activity: 84

Merit: 10

Good morning Infrastructure Committee,
This is a brainstorming session. We stand at the helm of the most efficient blockchain technology available. We need to set a course. We need a roadmap and timeline of Nxt goals.

My Proposed Goals:
Nxt AE release must be coordinated with an effective Marketing Campaign.
Nxt needs to (IMMEDIATELY) put together proper advertisements for programmer jobs complete with description of required duties and offered pay.
Nxt needs a USDollar to NXT registered exchange, preferably in New York State.
Nxt Point of Sale Use Case development.

We should open subthreads once the first top one or two priority projects on the roadmap are identified.

NXT Infrastructure Committee please brainstorm your ideas for Nxt infrastructure goals.

igmaca

full member

Activity: 168

Merit: 100

Quote from: opticalcarrier on March 16, 2014, 10:34:26 AM

Quote from: igmaca on March 16, 2014, 07:18:52 AM

exact that's the idea.

just trying to help nxt to be successful

what you are discussing can be done voluntarily now, but manual work and overhead required to accomplish your goal will be extremely high, and everyone involved will have to trust everyone else, as in a mesh: (n x (n-1)/2) links of trust. Actually, will even need to multiply that figure by 2 since the lines of trust have to be 2 way. This is why no one is responding to your suggestion - it will end up being a nightmare to manage manually, and just about as much of a nightmare to design/add/test the code for it to be done automatically by the network.

The way leasing will work in the very near future is that there is MUCH less trust involved, n-1. where just 1 way trust from each leasor to the single leasee. The leasee will still be responsible for calculating and returning the gains, though this *could* be enforced via code by the network, with a bit of work - certainly nowhere near as much work as would be required network enforcement of your shared idea.

I do get you point on it though, you get the psychological effect, then there are still more forging nodes out there; but unfortunately I dont think we will ever see it implemented in code as what you are describing will require tons and tons of work.

grateful to you for answering,

it seems very convincing your explanations.

if it can not be implemented easily by nxt not be easy by any clone

opticalcarrier

full member

Activity: 238

Merit: 100

Quote from: igmaca on March 16, 2014, 07:18:52 AM

exact that's the idea.

just trying to help nxt to be successful

what you are discussing can be done voluntarily now, but manual work and overhead required to accomplish your goal will be extremely high, and everyone involved will have to trust everyone else, as in a mesh: (n x (n-1)/2) links of trust. Actually, will even need to multiply that figure by 2 since the lines of trust have to be 2 way. This is why no one is responding to your suggestion - it will end up being a nightmare to manage manually, and just about as much of a nightmare to design/add/test the code for it to be done automatically by the network.

The way leasing will work in the very near future is that there is MUCH less trust involved, n-1. where just 1 way trust from each leasor to the single leasee. The leasee will still be responsible for calculating and returning the gains, though this *could* be enforced via code by the network, with a bit of work - certainly nowhere near as much work as would be required network enforcement of your shared idea.

I do get you point on it though, you get the psychological effect, then there are still more forging nodes out there; but unfortunately I dont think we will ever see it implemented in code as what you are describing will require tons and tons of work.

igmaca

full member

Activity: 168

Merit: 100

Quote from: ?? on ??

Quote from: opticalcarrier on March 15, 2014, 08:28:34 PM

Quote from: EvilDave on March 15, 2014, 07:04:22 PM

Does anyone have a brief summary of Igmacas idea(s), for the slow among us?

hes suggesting an idea but I cannot tell how he intends the network to enforce it

From what I can tell, there are two ideas in here:
* Proposal for a "share fee group" where accounts can group together (without central pooling) where any account that forges a block will share fees with group.
* Proposal to skim some percentage off of assets traded on AE as another fee.

My remarks on the matter are that we must wait and see fully implemented Transparent Forging before suggesting any radical changes to the core network function. I would expect smaller servers to be penalized by 100% TF for subpar performance -- "win or die". See these comments:

Quote from: Come-from-Beyond on January 28, 2014, 08:40:30 AM

Let me to interrupt u guys.
One of the next steps to 100% Transparent Forging will include changes that will lead to "grouping" of forging accounts. The main purpose is to make it impossible to predict who will forge the next block. If Alice, Bob, Charlie and Dan r next candidates to forge a block, then Alice, Bob or Charlie can do it, no matter if Alice would forge the block 5 seconds earlier than Bob or Charlie. The gap between blocks will be exactly 60 seconds and timestamp will be used only to determine an order of events. No sense to wait 93 seconds for the next block if we already know who would forge it. This is what I call "time warp". If Bob's block wins the race then Alice and Charlie will be penalized for 1440 blocks.

Quote from: Come-from-Beyond on January 29, 2014, 06:13:04 AM

Quote from: pandaisftw on January 29, 2014, 04:25:35 AM

Why is Charlie being punished? I assume the order of events is Alice -> Bob -> Charlie, so it makes sense Alice should be punished if she withholds her block and Bob ends up generating the block, but why Charlie?

Coz Charlie wasn't the winner in this race. Win or die, no half-measures.

PS: This is an example of dynamical equilibrium that heavily exploited in Nxt concept.

exact that's the idea.

just trying to help nxt to be successful

opticalcarrier

full member

Activity: 238

Merit: 100

im having a hard time with this POW pool thing. I have no clue on running pools, but I should be able to come up with some cron scripting to handle the API between the pool and coinwarz, and between the pool and an exchange. so if someone that has experience with pools wants to work with me we can split it.

Or if someone can point me to some pool server software. All Ive found is eloipool, a python based one, but is seems to come with no self-registration webpage, etc etc. so if anyone can point me in the right direction on this or wants to work with me let me know

opticalcarrier

full member

Activity: 238

Merit: 100

Quote from: EvilDave on March 15, 2014, 07:04:22 PM

Bloody hell, Ian and Igmaca.....

I must confess that Igmacas posts make my head hurt a little, and I may be unfairly ignoring him because of it.
My own personal time issues don't help, either.
Does anyone have a brief summary of Igmacas idea(s), for the slow among us?

hes suggesting an idea but I cannot tell how he intends the network to enforce it

EvilDave

hero member

Activity: 854

Merit: 1001

Bloody hell, Ian and Igmaca.....

I must confess that Igmacas posts make my head hurt a little, and I may be unfairly ignoring him because of it.
My own personal time issues don't help, either.
Does anyone have a brief summary of Igmacas idea(s), for the slow among us?

chanc3r

sr. member

Activity: 952

Merit: 253

THIS IS A THIRD POST FROM IGMACA FOR DISCUSSION

Quote from: igmaca on March 15, 2014, 01:24:18 PM

Thank you for answering my message.

It is very difficult for me to summarize the idea. I am not native Exnglish speaker

abstract

a) limit forging power to 0.1 (limit 1/3) total active balance maximun per account.
b) probability of forging success for any account and funds must be about one day (share fee group)

if you have 1.000 groups (share fee groups) and you limit forging power to 1.000.000 Nxt per account

then

1440 blocs per day means that 1.000.000 Nxt share fee group (many nodes) or one account (one node) forge every day
if you have more than 10.000.000 Nxt in for example 1.000 VPS nodes then 1.000.000 Nxt per account is less or equal than 0.1 * 10.000.000 Nxt = 1.000.000 Nxt

see attached file

http://www.docdroid.net/9yub/forging.pdf.html

Quote

It Should be Observed That this probability does not out stretch to 0 (asm! 1) if the bad guy has at Least 1 = 3 of all active balances in the network in the case of U-algorithm (corresondingly, At Least 1 = 2 in the case of exp-algorithm). There should exist some specific methods for protecting network against Such an attack in the CASE WHEN there is risk that (active) relative balance of the bad guy Could Become larger than the above threshold

it may be a good idea to limit the forging power of accounts by some fixed threshold, e.g., if an account has more than, say, 300K NXT,then it forges as if it had exactly 300K NXT. Of course, a rich guy can split his fortune between smaller accounts, but then all those accounts would forge roughly as one big account (without threshold) under Exp-algorithm. One can one use the computationally easier U-algorithm without having the drawbacks discussed in this section.

Quote

I don't get your idea, or I get it. I don't know.

Instead of leasing forge power, you commit with your account to share fees among others in the same "share fee group" if you forge a node. You still try to forge a block on your own, but you commit to share the incentive with others if you are successful (with special conditions like committing to run the node for some time, ...). Could this be done with AT and would that make sense?

chanc3r

sr. member

Activity: 952

Merit: 253

THIS IS A SECOND POST FROM IGMACA FOR DISCUSSION

Quote from: igmaca on March 14, 2014, 09:29:35 AM

Quote from: gck on March 14, 2014, 09:18:24 AM

This is my question as well. Instead of using nodecoin as the sole incentive for running a node...why not let the OTHER assets on NXT be the reward? You could reward people small amounts of any coin (or asset) they desire from the trading fees (small amounts...say 0.00000010 or less). Someone with a large stake in NXT could get slightly higher amounts. That way even someone that has just 1NXT...gets something for being an active node. That would give one hell of an incentive to get into NXT. Everyone becomes there own micro exchange.

Am I completely off base here?

of course! is a great idea!!!!

Everyone becomes there own micro exchange.

this is the solution. Nxt AE charges a small Nxt coin fee for each transaction and distributes
the fee to the active nodes

The active nodes receive two fees

A fee for each exchange assets independent of fund account
A fee as currently does nxt if node forges

if you want to use Nxt AE you have to use Nxt coins to pay the fees

Topic: [NXT] NXTInfrastructure committee - page 2. (Read 6344 times)