NRS Brainstorming Megathread

The easiest metric is total unique data stored (eg. Exabytes). This could be estimated from current section prefix length and average data storage per section.


The easiest, most useful, or both in your opinion? I could see that being actually more useful than unique addresses because of

  • deduplication (so it’s novel data)
  • you have to pay to put (so highly unlikely it’s spam to provide false growth metric)
  • shows actual growth of users/adoption, novel data, and network splits/growth.

Love that @jlpell great stuff


It’s a bit unclear to me what’s being proposed here.

Are we suggesting that there be a fixed number of NRS domains that can be purchased for 1 SNT?

And using that as the basis of SNT valuation?


Not really, more that we make 1 SNT buy a certain fixed amount of something in the first few sections. So folk holding SNT initially get a valuation based on network services for the network start. After say section 4 (prefix size 2) it’s just business as normal where the network “price” discovers.

Could be for network start 1 SNT == 100Gb of storage for instance.


Interesting, no idea if it is an issue or not but I wonder if the amount of storage chosen for 1 SNT isn’t right, could that lead to people either not storing anything at the start in the hope they can get it cheaper, or everyone trying to store everything they think they may ever want to store all at once and overwhelming the network because they think it will get a lot more expensive in the future?

Generally fixing prices centrally is difficult to get right, but I don’t know if that’s relevant to a temporary thing like this.

SNT should ultimately be related to the cost of storage on the Network as you say.

However, for NRS, I’d argue this is one of the few areas where it would be advantageous for the cost to be decoupled from the storage size.

The purpose of an NRS name is in making a XOR address human readable/pronounceable, with a name that’s selected by me; that’s where the value is.

So it would ideally have a price significant higher than the raw storage cost of the string, but not unattainably high. So perhaps a dollar or two rather than $300.


Yes, that makes sense.


Seems like the easiest way to represent that is to have NRS cost be at a fixed ratio to the PUT charge instead of a fixed ratio to SNT. For example, 1NRS = 1000 PUT. In terms of Fiat, the cost of a NRS reference then naturally tracks the price of 1GB storage.

If you want to make domain squatting more expensive then you could set the ratio at 1NRS = 1000000PUT to have a name track the fiat price of 1TB of Safe data storage.

Rough estimate - Based on current hdd prices per TB and 8x redundancy this would be about $200 per NRS at the high end. Otherwise you get sub dollar pricing for the 1:1000 ratio.


This concept has always been problematic to me. All NRS are not equally valuable.


To humans and that is subjective. So if one is less valuable and one knows it then they may not make it, if it’s valuable to them then they likely don’t care, and if it’s more valuable to others then they are making it to sell at an appreciated price.

Unless you have a proposal I’m not sure how else it’s dealt with. There have been discussions on domain squatting in the past that are worth a read but not so much on the cost I don’t think.

Regarding squatting, this was my favorite solution by @Seneca RFC - Decentralised Naming System II - continuous auction (by Seneca)


Mine too!

Well… Does that mean that the discussions about NRS possibly just being an addon and being exchangeable is from the table…? Because if not and you’re planning on making an expensive nrs system I’d be tempted to make an inofficial one that just charges for the storage costs

Just to state this: I’m pro NRS just being an exchangable plugin and I don’t see why there should only be one… I do want to try using a petname system too!

1 Like

This thread popped up in my feed the other day of someone’s first experience with web3, and an auction based domain sale. It’s worth a read.

Yeah, there’s nothing stopping you!

I’ve said it a bunch already (sorrynotsorry) but I agree a petname system would be great! But it would work alongside the NRS rather than be substitute for it.


I think what is different with Seneca’s approach is that registering is unique because it’s the domain + public key so there can be countless, therefore it can still be cheap and not a bidding war.

What the browser has to do is resolve the highest rank which is the “continuous auction” part, someone not getting PtD or without a community etc is going to have to pay continuously whereas a legitimate domain that is in use ought to be able to afford keeping a top rank. Doesn’t have to be excessively expensive either.

Plus would still work with the idea that 1 SN(T) can afford you 1B (or whatever number) of domains.

Flexible and a win win for sustainability and adoption to me.

Let’s keep this to ourselves though, eh? :wink:

1 Like

Imo, NRS should be practically free for almost all names. However, distributing current big names in this way will lead to squatting. We’ve talked to death about it, but £200 for safe://google will be an utter, total bargain. However, safe://1aszl3 would barely be worth a penny. Same number of characters, safe data storage cost, but completely different context.

For folks who just want a name it should be near free. For those who want a very special name, it is a completely different ball game. Balancing the two will be key to the success of NRS, imo.

I still think that maidsafe could hold big names and auction them off. It would fund the project with the money that some squatting chancer will claim otherwise.


Isn’t this not taking into account Seneca’s idea? Perhaps I’m misinterpreting and not doing or justice but if a domain is domain name + pub key to the network then can there not be x number of’s because each registered is unique? It just comes down to who has highest rank for the browser to resolve with close ties being presented to the user.

Seems like a good solution to me and quite unlike the above auction systems mentioned.

1 Like

And auctions like this are often peppered with people deliberately bidding just to jack the price up. When someone really wants something they will keep bidding higher and the trolls love pushing them up.

I agree since we wish to always have Safe for everyone, even those with little. And if each file I want to publish pubically has to pay bigger for the NRS name for each file works against people wanting to publish scientific/technical notes/tables/papers since they often fit in the 3 chunks but costs significantly more for the human readable name.

If NRS is a modest 10x the cost of a chunk store then each paper is costing 3+10 put costs for each table/note/paper rather than 3+1. A 3 times increase in cost. So if I were to publish a ton of tech stuff from the 60’s then its costing 3 times to do it.

But if NRS is 20x or 50x the cost then no way can they have a human readable NRS for them

1 Like

Would be cool to bid on NRS names, but then also the ability to change ownership of the container if someone wants to sell the NRS url.


Normal you publish many files on one domain instead of registering a domain name for each individual file though. At least on the web, but it’s an interesting ideas having the possibility of a unique domain name for a file. safe://neo/mytechnicalpaper.pdf vs safe://mytechnicalpaper.pdf


This will be possible, and so a reseller market and bidding on domains creates itself.

But if this is the default, then what does onboarding look like? I create a Safe, then choose a SafeID like @jimcollinson, and then an auction starts? How long does it run for? How long do I have to wait to get my name? A week? Then I get outbid at the end, and I’m back to square one.


How many jimcollisons are there in the world? How many will there be in the future?

Yup. This is why the proposed auction method is infeasible.