Topic: [XCR] Crypti | Dapps | Sidechains | Dapp Store | OPEN SOURCE | 100% own code | DPoS - page 214. (Read 804701 times)

Mal,

Did you see the posts on the coinbox we are coming out with? 

The unit will run 2 nodes simultaneously, allowing a user to forge other coins.  This makes supporting the XCR network cheaper due to income from another coin.

Mal,

The PoT algo is merely to give each and every node a turn at forging a block. irregardless of number of XCR owned.  There is a small boost to merchants based on their total sales.  With a 100 node network, the average guy will get 14 forges a day.  A merchant with several sales a day may get 17-20. 

Do you have a suggestion for how to make the forging equitable for all nodes, irregardless of stake?

Sounds like a great idea to me.  Why do you think this is bad?


PoT forging does nothing to address vendor Chicken / Egg.  Before the merchants show up, there's only "lottery winnings" to motivate a few forgers, not a stable system that rewards all forgers equally.

Viral spread of a cool app for electronic payment via smartphone using purchased "tokens" called Cryptis does.  Refocus from PoT to that app - it is what users will sign up for to download.  Start spreading it in Africa - they already understand and use this technique because they don't have banks.

Your assumption of "$3/mo" shared bandwidth and so only $60 per year as the total cost of running a node changes the number in my key point from $15K per year to $3K per year without changing my logic.  So here is my logic with your number:

Bottom line, if you assume $5 per month to run a node and specify a 2% forging fee per block, each and every newly added node needs to support an additional $3K in sales per year just to break even via a PoT forging reward for running the node.

In reality, I will back off of a $20/mo dedicated DSL line to your number of $3/mo for shared bandwidth when you show me a working every-node-to-every-node PoT algorithm that uses only 20% of a DSL line's bandwidth.  Just what IS the bandwidth requirement for a PoT algorithm that's got to communicate to every node once per minute the online status of every other node?



With all due respect, you are comparing apples and oranges here while using smoke and mirrors.

It costs money to put a node on the network: money for a computer, money for electricity, and money for bandwidth.  Today none of these can be bought with Crypti, so I have estimated their costs in US Dollars.  My original estimate was $300 per year to put a Crypti node on the network.  Your SWAG low-bandwidth PoT estimate was $60 per year.  That's the cost to start forging, estimated in dollars.  

To keep consistent, you've got to estimate the payback from forging in dollars, too.  It doesn't matter what the XCR to USD exchange rate is, it doesn't matter what the number of transactions in a day or a block is, it doesn't matter what toy XCR rewards get handed out as party favors before the transition to the true Crypti commerce marketplace.  All of these things are useless diversions that add nothing to the discussion or point I am trying to make.

Fact:  If the annual dollars earned from forging doesn't equal or exceed the annual expense in dollars for running the node, the forger is not motivated to continue.

Once the dust settles and Crypti is in routine operations mode, all that matters is what volume of commerce the nodes handle on a per block basis, measured in dollars since that's that's the currency unit that bought the forging nodes.  And my calculation stands:  For a 2% transaction fee, each node must support $15K (my dedicated DSL number) to $3K (your low bandwidth use number) in new commerce.  My logic stands for this statement.

Bottom line, letting somebody add a node without also requiring them to add an online shop is destabilizing the network by undermining its reward structure.  

The new forger wants a cut of the ongoing effort.  This cut has to come from somewhere.   It can only come from increased sales that generate increased fees.  A forger that does not add an online shop is expecting a vendor to generate the increased sales required to give the forger his expected cut.  This makes the forger a parasite, doing something the vendor already did when the vendor added HIS node as part of a new online shop.

This is a logical conclusion.  Open your mind to embrace its implications.  The key to Crypti success is vendors, not forgers.



I know you've been working for months on PoT.  It didn't work like you thought.  It's broken.  You're embarrassed.  You want redemption, you are calling in outside experts.  I get all that.  I understand you think PoT ***IS*** Crypti - BUT IT'S NOT.  The soul of Crypti is not PoT rewarding of forgers.  The soul of Crypti is a focus on commerce and vendors.  FOCUS ON COMMERCE AND VENDORS.


All you are saying here is that my calculations are correct of how many blocks per day a single node will forge in a 100/200/500/750/1000/3000 node network.  Embrace their implications.


Again, if there were no transaction fees and only vendor nodes as I suggest: (1) the vendors would make exactly the same amount by adjusting their prices upward slightly; (2) there would be no forgers, disappointed or otherwise; and (3) customers would be paying less in a more price-transparent system.

CryptiKit Update - Simple & easy Crypti node deployment and management.

New release tag for CryptiKit which fixes some issues I encountered when running the check_nodes task.

Please download here: https://github.com/karmacoma/cryptikit/releases/tag/v1.6.2

If you don't know about CryptiKit and what it can do already, please read here:
- https://github.com/karmacoma/cryptikit/blob/v1.6.2/README.md

---

Changelog

• Node's sync status is now checked before querying mining info or account balances
• Added warning message when a node's blockchain is still being loaded

Coming Soon

• Support for more distributions other than Ubuntu 14.04 LTS (64 bit)
• How-to guides on installing and using CryptiKit on various cloud hosting providers
• Optionally add passphrase to remote server when starting to forge

---

Cheers, Karmacoma24.

So focusing on these comments, this is something I also put quite a bit of thought into. In essence, the network would weed itself out in this manner naturally. Once it grew to a certain number of merchants, who were running their own nodes / implementation, whether it be through ATMs, Point of Sale terminals, etc, the network would start to adjust itself based on the RoI. People would only run a node and forge so long as it was profitable. We see this in Bitcoin mining.

Early adoption would be from individual users who would be supporting the network and forging, but over time it would adapt to be run by those who weren't concerned with RoI due to the declining rewards as more nodes enter the network and limit the amount of blocks each node forges. This, as you said, leads you to assume it would be vendors running a Point of Sale terminal. I had already considered the possibility and likelihood that this would eventually occur organically, but the initial adoption and user base is an important aspect of building and making Crypti successful in my mind.

Fundamentally speaking, if you are going to eliminate PoT and build the network irrespective and unconcerned with user base issues, why not just build a centralized network internally and then build it out through PoS terminals to vendors without ever involving any form of community in the actual running of the nodes? At that point all you would need is a strong web wallet with multi-sig, 3 factor auth on transfers, and added features such as a merchant directory and built in promotions, etc. You could focus the majority of your development around merchant outreach and the web wallet.

You still run into the same issue with early adoption. Merchants won't sign up if no one is there to spend it and users won't sign up if they get nothing in return. Chicken / Egg.

@MalReynolds

I completely understand what you are trying to say and don't necessarily disagree. It is something I have put a lot of thought into myself. However, if you were going to develop a low transaction fee financial payment network, that was geared towards merchants and offered the benefits of typical crypto currency markets (i.e. fast transactions, low Tx fees, etc), how do you entice USER adoption without some form of reward? Do you think that PoP rewards of transactions fees (essentially cash back) is enough motivation?

Merchant adoption would be a simple matter of getting the ball rolling with free point of sale terminals and setup for some businesses to test the market, assuming you have a payment conversion process setup to allow them to receive payment in USD (which is still unfortunately necessary).

USER adoption on the other hand is more complicated with out built in incentives. Bitcoin gained user adoption because they could mine the coin and profit from it initially with GPUs that they already had on hand which made it intriguing.

The bottom line is, if all you are offering merchants is a low cost payment processing and point of sale system and you are offering consumers relatively nothing, why wouldn't they just accept Bitcoin (which already has a built in user base) and call it a day? What motivates them to switch to Crypti?

Interesting essay Mal.  I have a few comments:


the key cost is bandwidth at $20 per month

This assumes that a forger will have a DSL line dedicated to just the node.  I doubt that...... most if not all forgers will use the same DSL or work T1 lines that they use for all other internet communications.  Merchants wil already have DSL lines set up for their credit card transactions.  Homeowners already have their DSL set up, so a node connection presents no additional bandwidth fees.   That lowers your estimate of $300 a year to $60 a year.


There's only 1440 Crypti blocks to be forged in a day

]FOR A COMMERCE BASED CRYPTOCOIN THAT USES FORGING AS A REWARD, THE NUMBER OF NODES MUST BE KEPT IN LOCKSTEP WITH THE INCREASE IN SALES REQUIRED TO SUPPORT IT


But you assume that the only transactions going thru the node for the next year are the Cryptsi Node Reward Program transactions.  There will not only be other transactions, but each block can contain 254 transactions.  So we ar elimited to 365.000 or so transactions a day.  A reward for forging 1 million XCR is 5000XCR.

Today there were ten blocks that went thru with 100XCR fees, and one with a 1000XCR fee. That was the devs making a merchant account for the Node Reward Program.   Lucky Forgers.

Additionally, you are also assuming that XCR will stay the same price relative to the USD as it is today.  I doubt that very much, and so do our frequent bloggers.

For a merchant to become a verified merchant, they must apply and pay a 1000XCR fee.  Then 10 forgers must verify the merchant.  Those 10 forgers earn 100XCR each. 


Of course, if you were to DROP PoT..

PoT is the backbone of XCR.  The team here has been working on the specs for PoT the past few weeks.  The current plan calls for it to be the main algo to determine what node forges a block.  The other algos are just frosting on the cake, adding a little to the value, but not a lot.

PoT will not allow a long-time up node to dominate the forging.  Each node is given one point for every block it is up, and once it forges a block, the PoT value is reset to 0.  That allows the other nodes to in the queue to forge, and prevents a type of attack.

The other value for merchants is the sharing of the transaction fee.  Even your wildly estimated cost of $300 a year ($1 a day maybe) can be made up with just one daily earned fee from a $200 sale.

The whole point of PoT is that (1) everybody is motivated to keep a node going as long as possible and (2) everybody who HAS kept a node going has an EQUAL CHANCE to forge a block whether you are a whale or a minnow.  Can PoT forgers be rewarded enough to keep them motivated to forge Crypti?  Let's explore that question.

First, how much does an always-on personal Crypti node cost?  Let's go with a $75 2W PCduino amortized over three years using $3 per year of electricity (17.5 KWh per yr at $0.15 per Kwh) and $20 per month for a continuous DSL line running 0.75 to 3 Mbps for bandwidth.  The computer and power is thus only around $3 per month; the key cost is bandwidth at $20 per month.  Let's say a personal always-on Crypti node is thus $25 per month or $300 per year.

There's only 1440 Crypti blocks to be forged in a day.  Roughly speaking, given the EQUAL CHANCE aspect of PoT:

Your single node in a 100 full-time node network captures 14 blocks per day, 5110 blocks per year; breakeven requires $0.06 per block
Your single node in a 200 full-time node network captures 7 blocks per day, 2550 blocks per year; breakeven requires $0.12 per block
Your single node in a 500 full-time node network captures 3 blocks per day, 1095 blocks per year; breakeven requires $0.27 per block
Your single node in a 750 full-time node network captures 2 blocks per day, 730 blocks per year; breakeven requires $0.41 per block
Your single node in a 1500 full-time node network captures 1 block per day, 365 blocks per year; breakeven requires $0.82 per block
Your single node in a 3000 full-time node network captures 1 block EVERY OTHER day, 182 blocks per year; breakeven requires $1.64 per block

So forging fees for a 100-3000 Crypti node network needs to be somewhere between $0.06 and $1.64 per block JUST TO BREAK EVEN with PoT forging.

Let's say that the forging fees are adjusted so that they are around 2 % of the total amount of transactions going across the Crypti network.   ATM and credit cards fees are around 3%, so this makes using Crypti a bargain compared to using a plastic card.

A 2% fee means to get $0.06 in forging fees per block, the AVERAGE block must have 50 times that much in sales transactions, or $3 per block.  That's Crypti sales of $4320 per day or $1.57 million per year.

A 2% fee means to get $1.64 in forging fees per block, the AVERAGE block must have 50 times that much in sales transactions, or $82 per block.  That's Crypti sales of $118K per day or $43 million per year.

So just to break even with a competitive 2% forging fee, a 100 node network needs to have annual sales of $1.57 million or $15.7K annually PER NODE.

And just to break even with a competitive 2% forging fee, a 3000 node network needs to have annual sales of $43 million or $14.3K annually PER NODE.

This $15.7K - $14.3K range is really a single figure per node regardless of network size, the difference is due to roundoff error in each scenario.

Bottom line, if you assume $30 per month to run a node and specify a 2% forging fee per block, each and every newly added node needs to support an additional $15K in sales per year just to break even via a PoT forging reward for running the node.

Now, here's the gotcha - any person running a node that generates $15K per year in sales is not a "miner" or a "forger" - they must be a VENDOR by definition.  A vendor that must add their own (always-on) Crypti node onto the network to allow processing of Crypti as a currency in a transaction.   Any guy who just slaps a new PoT forging node on the network WITHOUT SETTING UP AN ONLINE SHOP TO GO WITH IT is a PARASITE that is requiring the TRUE vendors to sell (MUCH) more and so keep the average sales-per-block figure high enough to support the forger.

Note how in order to keep a reward forging network adequately funded, the size of the network must scale in lockstep with the sales revenue it is supporting.   If people become forgers at a whim "just because I want to forge some XCR"  or "just because it makes the Crypti network stronger", then the economics of the network are DISRUPTED.  

FOR A COMMERCE BASED CRYPTOCOIN THAT USES FORGING AS A REWARD, THE NUMBER OF NODES MUST BE KEPT IN LOCKSTEP WITH THE INCREASE IN SALES REQUIRED TO SUPPORT IT.  

Just how does Crypti propose to do THAT?

Of course, if you were to DROP PoT....and DROP forging...and JUST GO WITH PoI / PoP... then a CUSTOMER pays NO visible transaction fee AT ALL (yipee!)...a VENDOR could add a node anytime they wanted, and write off $30 per month as a cost of doing business in running their own node AS A PART OF THEIR ONLINE SHOP, and they wouldn't have to hit a target of $15K in sales per year (or more!) to guarantee the fees are there for some forger to stay motivated...  

Eventually you guys will acknowledge my point and realize it's VENDORS, not FORGERS, that are the key to running a successful Crypti network.   And on that day, you will drop this crazy insistence on perfecting PoT to motivate forgers get on with making Crypti the key cryptocurrency for commerce...

When I said "blockchain", I mean the node tab "blockchain".  It should show the latest block and the current tine. 

I checked mine 30 mins ago, and found it had stalled just when some large transactions when thru and some users got 100 XCR. 

I have installed the same rebooter program I use on my miner.  I have been rebooting my miner every 12hrs automatically with this program.  Now I have it on the node computer as well.  www.passmark.com if you want it. " rebooter".  You can set it to any time frame you desire.


BTW, once you have enabled forging, it will forge after a reboot without you entering the passphrase.

I checked my windows wallet for blockchain time and it was not showing the current time. Why does the wallet shows 'Forging Enabled' if it
not really forging ? Not cool.

So you are the lucky one........

We are setting up two wallets for the Crypti Node Reward Program. 

There were 40,000 XCR sent to one of the wallets.

Hope to have the program up and running real soon.

while we were looking at the windows node it appears my t2.medium linux box forged 281 XCR:



Somebody correct me if I'm wrong.

~informer

I'm positive it's forging.



I was only looking at the nw.exe. Yeah, in the Resource Monitor it shows 3 node.exes's and they're using about a GB of memory. thanks for the info Lit.



~informer

20-30MB or RAM?  sounds like it is not really running.  Look at the blockchain page.  Is it current to the correct time?  The node can stall, and still the window will appear normal, but no updating of the bc.

Check your Resource Monitor in windows.  Under the memory tab, there should be 3 "node.exe" running.  One of them should be around 700MB.  

If you only have 2 instances of "node.exe" you are not forging.

+1

~informer

Thought I'd share some anecdotal notes:

I'm running one Windows Wallet on an old core2duo with 4 GB's RAM from the house. I have some QoS rules in my router which allows me to prefer traffic to its IP. I don't know if it helps or not as I have not had the time to do some real testing. It's been up for over a week now without restarting or rebooting. It only uses about 20-30MB's or RAM and never goes over 5% CPU usage. Same goes for a AWS t2.micro running Windows server 2012 R2 Standard. Same performance essentially.

However, it seems if I want to run a linux box, I have to run it on a t2.medium which has two proc's and 4 GB's of RAM. On the medium Linux box, crypti uses about 2.5 Gb's of RAM. If I run it on anything smaller (on AWS) it will fork due to out of memory issues. I also have several instances on Digital Ocean that exhibit the same behavior.

~informer

When it hangs, restart the computer, not just the node.

I now use an old laptop as the node, with the node  as the only program running, other than the other 40 window's stuff.  It does hang after 4 or 5 days, so then I restart it.

The devs are working on the mem problem, as related in a post two days ago.

Oh, now i cannot run Crypti even for two hours, not days. With enough memory it hangs and not recovering. 
Seems this 1 coin for block reward is not for me 

I did not know about this.
Awesome idea to expand the network. Good luck today.