Hello again, and a few questions!

Community
#1

Hi all :wave: I used to be a frequent user here back in 2014. It’s sad to see the network hasn’t been finally released after 7 years, but I’m happy that this project hasn’t been abandoned, as it’s now needed more than ever. I hope SAFE still has a successful launch!

I’ve been out of the loop for a while and have a few questions for you:

  1. Are my MAID coins still safe? :sweat_smile: I originally moved them to my own local wallet when the original exchange went under (Poloniex?) The coins haven’t been moved to another blockchain or anything, right?
  2. What was the final decision with Pay the Provider (PtP)? Is it still a consideration?
  3. Can you clarify what exactly still needs to be completed before a “final” release, with real SAFE tokens? The roadmap doesn’t make it clear. It says Safe Maxwell Network is coming next, but it only “sets the ground work”. Then there are Apps and User Experience items – do all those need to be finished first as well? A more concrete list of what’s required vs what’s nice-to-have would be great, with a final goal of “RELEASE”.
  4. Technical question: I see that adding data to the network is pay-once-store-forever by default. What incentive do users who provide storage have to store forever? Do they continue getting paid over time, and if so, are tokens just created out of nowhere? And if not, why should users store data for years, that they’re not getting paid for anymore? Especially considering churn, where users may get old data transferred to them.

Cheers!

11 Likes
#2

  1. They are still on the same blockchain. There are some community efforts to get a wrapped MAID that can be traded on decentralized exchanges, but that hasn’t come to fruition, yet.

  2. PtP is still being discussed and the convo was just resurrected a month or so ago. No final decision.

  3. The main thing that is left is in place upgrades, and getting farming algorithms nailed down. Nearly everything else should be in the upcoming test network that we believe will be out this coming week (with bugs and known issues, etc., it is a test network).

  4. Farmers will churn to other sections as sections split, etc. There should always be chances to earn.

5 Likes
#3

Hey Ephi, great to see you again.

  1. Yes your MAID are still good.
  2. PtP is still a maybe, no change there.
  3. The test net coming this week will be feature complete except for upgrades and maybe one other thing. So check that out and maybe read a couple of recent dev updates (there was none last Thursday, but that’s a rare exception). There’s a Safe Network App and the Browser which need updating, along with APIs, but all that is also ongoing in the background so very little totally new to do. Quite a bit of testing and tweaking, eg of reward/economics, but features are pretty much done.
  4. Nodes get paid for holding data, and other “correct behaviour” so of they neglect their duties they get punished, paid less or kicked out and have to earn their age again.

See you around.

7 Likes
#4

Thanks for the info @wydileie and @happybeing, and hello again @happybeing, good to see you’re still here :grinning: I’m really glad to hear the network is almost there!

Nodes get paid for holding data, and other “correct behaviour”

Can you please clarify this? If users only pay once to store their data, where do ongoing payments come from?

Tangentially, what kind of total token algorithm is planned? Will there be a total cap (like bitcoin), infinite linear increase, or something else?

1 Like
#5

There’s a maximum number of 2^32 tokens.

Nodes receive a payment from the section when it splits, an amount proportional to their age. The section uses payments accumulated from storage plus newly minted tokens to fund this - this is the plan for the test net at least. Certain bad behaviours cause age to halve, others get the node kicked out so it starts at zero.

2 Likes
#6

So this is a change from being paid for GETs? Or is it in addition?

#7

Right now, it is different, we penalise bad nodes and pay for behaving (and storing) totally on split. This is to get feature complete testnet though. After that it’s all open for debate. i.e. a node does not provide Get requests will be killed. So not fastest or slowest, but in the acceptable range.

Right now you need to store to Give or be killed. So not payment per Get.

5 Likes
#8

Hello and welcome again Ephi!

In the Bitcoin network, when a new node join, it receives the entire blockchain. It does not receive payment for this. It receives payment only for future work.

Safe works on the same principle. The assumption on which the network economy is based is that internet speeds and storage space will grow along with technological progress, and if storing 1 GB today for 10 years costs $ 1 (for example), then in 10 years the price will have dropped to 1 cent. Ie the new nodes on the network will receive old data that currently costs them a negligible amount to store compared to what it cost years ago.

In summary, the user pays the network the full amount in advance, but the network pays the nodes little by little over time, and as technology evolves, for the network is necessary to pay less and less for the same amount of data to the nodes. Ie in theory, the art would be to hit the farming algorithm to be in little advantage for the network over time. :dragon:

Privacy. Security. Freedom

5 Likes
#9

I’m not sure what a “split” is so I’ll have to read up on the primer.

the user pays the network the full amount in advance, but the network pays the nodes little by little over time

I think this answers the question pretty well.

the art would be to hit the farming algorithm to be in little advantage for the network over time.

As in, to make the price to PUT a bit cheaper than the price to store?

1 Like
#10

Sections have a minimum of (currently) seven and maximum of twice this number of nodes. They add new nodes as they need to store and manage more data, and so split in two when they get to the maximum size.

Data (chunks) are stored in a specific section according to the XorName (address) of the chunk and section, each section having a prefix which matches the data XorName by one, then two then three etc bits as the number of sections grow.

At least that’s what I understand. I think the primer is reasonably up to date but haven’t read it recently to be sure.

3 Likes