Wishful Thinking about "Alternative" SAFE Networks

Related Projects
1

I’ve thought about these threads for awhile.
Using the SAFE network without money
Safecoin divisibility
Path of Least Resistance
Final few points about PtP
Safecoin computational/communication density

Finally, I decided to brainstorm alternative features. One member said I will be crucified for posting this… He is probably right. As the title says, this is just wishful thinking.

If I were to build my ideal Decentralized Network, what would it be like?

  1. Zero barrier to entry… Free to upload/download.
  2. Turn my computer into cash… work from home.
  3. Has tons of free content… while others require payment.
  4. Currency accessibility… you earn what you bring to the network.
  5. Very efficient Network… requires very little processing.

Brainstorm Alternatives

  1. The Network doesn’t charge for uploading/downloading AND it pays out … SAFE GB!
    What is the purpose of SAFE GB? It’s used for “permanent” storage. This means all initial PUTS will be temporary (LRU - Least Recently Used Cache) until it is made permanent by paying SAFE GB. This allows people zero barrier entry to try Network without caps, limited trials, bait and switch gimmicks.

  2. I like supply/demand adjusting Network payout. So let’s keep that. When the Network is very empty, it pays out less SAFE GB. When the Network is nearly full it pays out more SAFE GB. Since there is no supply cap, the Network never has trouble paying out more to meet demand. This allows it to continue forever, but it’s still limited by productivity. At the same time, the currency is consumable from #1… permanent storage. Hopefully, this will help mitigate rapid inflation.

  3. I really like the walletmark idea, and believe it assists all payment situations. Definitely a keeper. Content is free to upload/download because of #1. Therefore, I think more people will upload content with walletmark, making it permanent if it becomes popular. Or the Network can charge for a walletmarking service, which includes: LIKES, TIPS, PAY on GET. Otherwise, it goes away as temporary content. Good for the Network, good for the producer, great for the consumers!

  4. IMO, the most annoying thing about currency is accessibility. A great selling point of Bitcoin was you could farm it on your average computer. But that is impossible now. I would like a currency that can be easily acquired today as well as 10, 20, 30 years from now. My only requirement is that productivity be the limit for currency creation rather than an arbitrary cap. The more resource you provide, the more SAFE GB you earn.

  5. Currency security is a big deal and it’s very important. But at the same time, forcing the Network to count individual transaction units is inefficient. A ledger account system would be better, we just need a larger consensus group or some other way to beef up security. Less processing for the Network makes it more efficient.

This changes some things while keeping others. It’s definitely not a perfect alternative. But it does sound more user friendly. I’m sure many will disagree on various points and I’m okay with that.

EDIT (UPDATE)

Thanks to many posts below, I have reconsidered some of the brainstorm ideas above. Also, changed SAFE GB to POR… (Proof Of Resource). It is a poor man’s currency. But one that is accessible world wide.

  1. Temporary storage is more irritating for new users. It would also require more maintenance by the Network to: track, remove, and toggle temp data to perm data. I’m scrapping it. All data is permanent and costs zero to upload… when storage capacity is less than 20%. But as storage capacity increases beyond 20%, PUT costs increase to discouraging uploads, while encouraging farming. See Q&A “Storage Overload.”

  2. POR is rewarded on a 1:1 ratio. If a vault successfully supplies 1 GET (1 MB chunk beating the other 3 vaults) they get 1 POR. This basically means farmers generate rewards according to the amount of “good” resources they provide, no more, no less… forever. “Good” means their GET replies beat the other 3 vaults.

  3. I think walletmark is perfect as is.

  4. Because of #2, POR will always be accessible by everyone, adjusted by current technology. If a faster delivery vault is created, that farmer will gain an advantage until everyone catches up. At the same time, this shouldn’t exclude lower tech farmers from earning something, because their reward is based on “good” resources provided, not limited by a cap.

  5. This one is a pickle. If ledger systems are less secure than individual data units, then I’ll have to settle for POR = 1 MB as the lowest denomination per data unit. Possible ledger solution see this post.

Since POR is naturally inflationary, it’s unlikely it will need smaller denominations. However, I could create “change machine” feature. It would burn POR and create a ledger balance. POR data units are classified as “Savings”, while POR ledger balance is classified as “Checking”.

Q&A

Why bother rewarding POR if has no purpose in the Network?
I see it more like a job… Your vault works 8 hours and gets paid for the amount of work it does. That’s it! K.I.S.S. What the farmer does with the reward is entirely up to them. They can trade it, transfer it, transact with it… etc. The currency remains relatively stable because, 1 POR will always require 1MB of “good” resources to generate, now and in the future.

What about spamming?
I think this will be an issue even with a pay system. I still get flyers in my physical mailbox. A better solution would be to subscribe to a “Do Not GET” block list. The spammers don’t know you’re not GETTING their spam, since it is being ignored on your end. Besides if storage ends up being as cheap as we expect, then marketing money will flow into AD spamming anyway.

What about storage overload?
I was one of the first to raise this issue. Here’s my solution.

Bootstrapping Solution: the Network adjusts PUT costs at certain thresholds.

  • 0% to 20% capacity, 0 POR buys unlimited PUTS.
  • 21% to 89% capacity, 1 POR buys X PUTS.
  • 90% to 100% capacity, PUT unavailable.

The reason we stop allowing PUTS at 90% is because we need a 10% reserve to compensate for a churn events that could result in data loss if there is no storage available. Some would argue 10% reserve is too little. But as the Network fills to 90%, the PUT costs have been increasing to reduce uploading.

Why 20%?
Because at 20% we know POR exists, and can be used to pay for remaining storage capacity. This solves the bootstrap problem, and remains flexible enough to adjust to overwhelming PUT demand. Vice versa, if the community contributes more than they consume, the PUT costs will remain zero. The Network reacts to community behavior.

What about inflation?
This is a curious thing because “bad” inflation is an increase in the money supply while the goods and services remain the same. This results in higher prices. The Network increases supply as a consequence of servicing a demand (GET requests and Storage). Keep in mind it’s not that easy to generate POR. There are already several factors that reduce payout: caching, other 3 vaults, bandwith and actual GET demand… all of which help limit currency generation.

But if the Network gets mass adopted, won’t there be trillions of POR floating around?
Absolutely, global distribution is the goal. I’m confident we can find ways for the Network to consume POR by providing useful service to end-users. Otherwise, people can use POR for everything beyond Network services.

Consumable POR
Spending this currency on core Network services consumes (destroys) it. Without a supply cap, the only way to remove currency from circulation is to use it for Network services. This gives it a purpose while countering inflation at the same time. What Network services consume POR?

Just to be clear, this is not about changing the current plan (Safecoin Incentive Model). Many of us invested in MAID, including myself. I like to write down my ideas for the future, in case they come in handy.

SD (Structured Data) Shares?

I had a discussion with @dallyshalla and proposed an idea that might be very useful. 1 Structured Data unit holds the previous and current owner’s public ID. What if we use the “Safecoin wallet address” as the owner ID?

How is this useful?
I create 100 Shares (Structured Data). And each share represents 1% of the company’s profit revenue. So if the company earns 1000 Safecoins, it pays 10 SC to each share’s wallet address.

If the owner of that share sells it on SAFEX to someone else, then the wallet address changes to the new owner. The company DAO looks up each share’s wallet address for Safecoin payout.

Is there an easier way to do this?

Is the owner’s ID the same as the wallet address?

Can it be that simple?

4 Likes
2

sigh

1)

A)

Time and @dirvine is like oil and water. That shit don’t mix. I am in agreement with him - and I’m sure that I’m not alone.

If you don’t mind believing that it changes everything,
Time will never matter
– Jars of Clay - Sunny Days

B)

You talk about operating the network without bait and switch gimmicks, then say that you can upload data, but you have to pay to make it permanent? Sorry, that sounds like bait and switch to me. (at least coming from the current paradigm on the existing internet)

C)

One of the advantages of P2P decentralized storage-based networks (IPFS, Freenet, etc.) is that they don’t delete data. I hate broken links.

D)

From the first sentence in “1”, I initially thought that uploaders and downloaders would get paid for uploading and downloading. If this is not the case, skip this point.

You can’t reward people for consuming the resources on the Network. This is a classic Tragedy of the Commons. Yes, it will deplete resources. No, it won’t advance the economic integrity of the Network.

I’ll just leave this here.

2)

A)

Since there is no supply cap - the payment value will be arbitrary. It makes no sense to have an unlimited supply pool of one thing (SAFE GB) while having a finite supply pool of the related, other thing (storage space).

value = (storage space) / (SAFE GB)
value = (finite) / (infinite)
value = lim(x -> infinity) of (finite/infinite) = 0

Read as:

The limit of a/x as x approaches infinity is 0
– MathIsFun.com

For this reason I find the name “SAFE GB” to be extremely misleading.

B)

If the payment is arbitrary this leads to inflation. Always has, always will. It actually follows the same proof as above.

C)

In a balanced network (and that’s what we’re building, right?) the algorithm (tuned correctly) will ensure that the Network never has trouble paying out more to meed demand.

I’m actually flabbergasted that you consider this a solution. To what problem? In the existing blueprint:

  • Network has little space = there has been much storage purchased = large amount of Safecoin in pool
    • Therefore larger farming rewards are in order.
  • Network has much space = there has been not much storage purchased = small amount of Safecoin in pool.
    • Therefore smaller farming rewards are in order.

If it ain’t broke…

D)

How do you plan on setting a price/reward for permanent storage space? Here’s your answer: it’s arbitrary. You have two options (both - once again - arbitrary):

  1. Fixed (magic) amount (let’s see how well that goes over with the community)
  • Deterministic amount

Well, what could the deterministic amount be based on?

  1. GETs?

    Aren’t the number of GETs not dependent on any real cost factor? They do not give monetary value to the network. They also cause more work to be done per monetary reward unit to the farmers. If this route is followed, although it may be free to GET, the more GETs occur, the higher the cost to perm PUT. That does not encourage people to use the network.

    We’re bait and switching people saying that viewing the network is free - but that the more you view, the more it is to perm PUT your data.

  • Perm PUTs?

    If you perm PUT less, it’ll be cheaper. If you perm PUT more, it’ll be more expensive. I don’t see that encouraging people to perm PUT their data. Also, how could you determine price based on the decision to pay that price? That would infer either a:

  • Free market - farmers set their price for space. The network doesn’t actually doesn’t factor into the equation at all. This has the strong potential to break the equal distribution that the data must necessarily have to ensure loss prevention.

  • Feedback loop - one relies on the other which relys on the first. That’s a catch-22 if I ever saw one.

  • Pool?

    Oh wait, there is no pool.

  • Farming payment amount?

    As shown above, that is arbitrary by definition.

E)

In the existing model the currency is still consumable from the very beginning - basic storage.

3)

A)

If having content become popular is the motivation to make people purchase permanent storage, what’s to say that they won’t just keep uploading it every ${temporary PUT time}?

B)

Regarding re-uploaded data: The hash will still be the same. The location on the network will still be the same. It would take me a day (two max) to script a cron job that downloaded my data 1 ms before it’s deleted, and then re-upload it 1 ms after.

C)

If my popular upload’s temporary timeframe expired before I was able to know that it became popular, what’s to stop someone else from uploading that piece of popular data with their walletmark?

D)

Even having a delete functionality in the network (not to mention an automatic one!) is dangerous to begin with. Bugs happen.

E)

Stop with the “Pay per Get” disillusionment. It’s an ill-conceived notion that is destined to bring down this network before it even begins.

4)

A)

Farming temporary data is analogous to the US’s Subprime mortgage crisis - except worse. You never know if that data will be paid for. At least banks have legal fallbacks if a debt is not paid.

You don’t see many programs offering free rent in the hopes that the place will be liked so much that it will be paid for.

B)

The system already has in place measures to ensure that 10, 20, 30 years from now farming will still be viable for the average consumer. The idea of Safecoin Farming Speed versus the vault size was detailed in the whitepaper: Safecoin: The Decentralized Network Token.

5)

A)

Ledger-type currencies seem to have scaling problems. 'Nuff said.

B)

Security is a big deal. We’ve seen that ledger-type currencies have enormous security and anonymity shortcomings. Refer to Andreas Antonopolous’ Q&A on the subject for more info. (Video linked to begin at relevant point)

Dude, I know that it’s Thanksgiving Eve, but ease up on the drinks, will ya?

2 Likes
3

Looks like he was right… :smiley:

4

I like the expiring idea and your intention (low barrier) is straightforward and important. I think gameability is again the main problem, but also technical problems (which I can only guess). Let’s say I upload my stuff and it expires after 30 days or 6 month.

  1. what keeps me from writing a simple script that renews files after 29 days or 5 months?
  2. how can I make files permanent since I am not directly related to the files (for security reasons)?

A possible solution: one could work with deduplication - so if content is reuploaded it doesn’t receive free days and you have to pay to save files. Usually this wouldn’t affect personal files. However, you can still use it as an attack vector.

I general I believe the network is sensitive and should be K.I.S.S.-designed without gadgets that constantly break. Anyway, I agree with you that there is a barrier and keeping it low will be decisive for the network. I don´t see why the first tools user will use (i.e. a browser / plugin) shouldn´t lead them to a faucet where they can receive free SAFE to PUT stuff (remember, they are free to browse as in oldschool internet). I think the way faucets currently work demonstrate the problem that you are dealing with when giving away ressources for free pretty well, so I think that´ll be challenging - and if we don´t find an appropriate way in solving it, this underlines that we shouldn´t put something in the code that messes with the grand economy of SAFE.

1 Like
5

I can just see Janey uploading her family photos for a test (not public), but her mother already uploaded those (and more) 29.5 days ago. The next day, Janey says to her husband in a proud way, see hubbie I uploaded all our photos in a secure place that no one else can get them. Then she proceeds to show him, but only a handful remains. The expletives, the swearing off using SAFE ever again.

No dedup could not be used to disallow the 30 days “trial”

I think the payment model actually mitigates again a lot of the potential problems that this thought experiment explores. But glad someone presented it because it brings together in one coherent way many discussions over the past few weeks/moneths

3 Likes
6

I like all but 5 - digital property in the fullest sense is an exciting prospect.

1 is a good idea too. Perhaps having an expiry date on PUTs would work, while retaining security anonymity? Vaults could then shred obsolete datamaps (data would have to persist due to security). However, as the data would persist regardless, other security concessions would need to be made, which I am less comfortable with.

Maybe there could be a mechanism to define whether security or price is ideal for an account?

1 Like
7

So the point is to try to make the barrier to entry as minimal as possible. With the existing setup, we’ve gone one step further. Not only the barrier to entry is minimal, but the ongoing use of the network is minimal.

90% of the internet is comprised of lurkers. And this is with the free “PUTs” that the internet has now. Surfing is free on the SAFE Network - so we don’t have to worry about that too much.

Anyone can “demo” the SAFE Network. When I go to an existing project, the first thing that I do is download the client and look around. Just look. That is entirely feasible with the Network as it stands now.

To actually use the network (PUT), one must contribute to the network. It’s like an actually fair bittorrent system if we want to put it to an analogy. This too has been made as painless as possible, with the most bang-for-your-buck imaginable.

Good point, and it underlines another as well. If there is another way to get Safecoins without contributing to the network (TIPS from content on the network notwithstanding as that is - although less tangible - providing a contribution to the network) the we can solidly say that the network is broken.

IMO the network is already designed to be as n00b-friendly as possible, while still maintaining it’s elegant simplicity. You can’t give something away for nothing in return - and expect to be around for very long that is. Anyone who wants to give out free storage and free coins feel free to send me a link to your safe-site/app. And thanks for all the fish.

Also as an aside, as much as I would want to believe that SAFE GB (in any of it’s forms proposed by @dyamanaka) would only be used for storage, that won’t be the case. If history has taught us anything, it’s that anything that can be used for currency - will be used for currency.

3 Likes
8

Timber and nails for sale! Timber and nails for sale!

3 Likes
9

Oh, c’mon. I wasn’t that bad.

EDIT: And for the record, I hold @dyamanaka in the highest regard. His opinions and suggestions are always put forth in good faith, and often substantially contribute to the growth of our collective understanding and the general evolution of the Network.

3 Likes
10

I have my eyes on the future mate, you’re not going to be the only one! :smiley:

1 Like
11

Excellent video (direct link) thanks for that, a must watch:

“…solving the anonymity and privacy issue [for bitcoin] is something we need to solve before the protocol ossifies because we won’t get too many chances, otherwise we will succeed in disrupting the financial services infrastructure with bitcoin, and a decade from now we are going to need to disrupt bitcoin with something else…”.

Offtopic: It can be argued that bitcoin has already ossified and it will be very difficult to tack on core privacy features at this point, so… the decade is waiting for Safecoin already :smile: . My concern (still) is this mistake might be repeated by leaving the Safecoin divisibility issue until after launch, that may also be too late:

  • -ve : Early on when the network is small it possibly can’t handle division transaction load so division is not included.
    • +ve : Few players so the core code can still be changed easily by core Maidsafe team without needing consensus.
  • +ve : Later on when the network is large it can handle division transaction load
    • -ve : Many players paticipating in larger network means the core code/protocols cannot be changed easily without requiring consensus. The network has “ossified”.
3 Likes
12

DYamanaka strikes the nerve again :thumbsup:
Let’s see if your topic will be moderated into closing like mine (in which I suggested free use for a short while).

  1. One problem (currently) is data can’t be deleted, so deadbeat users would leave their droppings on the network. (Secondly, if those droppings are public files, then the old problem of cheaper (here free) public data would be quite serious). This weakness applies to my “free” plan as well.
    Perhaps your point 1) would be more practical if it didn’t apply to public data (i.e. non-free public data and free private data)?

Tweaking may be (and probably is) necessary, but with everything already significantly delayed, I think it’s better to not make changes at this point in time.

If I had to choose between a non-tweaked (current) approach and completely free use (initially, while the tweaking is going on) by April 2016, i’d choose the latter.

1 Like
13

Bitcoin has always had poor accessibility. Sure, hardware-wise the accessibility was good in the early days, but then in terms of knowledge and skills the accessibility was very low. Only a very small group even knew about Bitcoin, and for the non-techies that by chance did know about Bitcoin, setting up a mining rig wasn’t easy either. In the end, it’s impossible to shower the world population with a currency without proportional inflation. Early BTC miners were rewarded for their scarce knowledge.

So you want to increase liabilities when the network is becoming insolvent? That seems suicidal to me. With SafeCoin it is possible to do this because the amount of storage a SafeCoin gets you can (and should) (temporarily) lower proportionally (or more) to the increase in rewards, i.e. desipte issuing more SafeCoins, net liability doesn’t increase or even lowers, until equilibrium is regained. I’m assuming a SAFE GB is an actual promise for a GB here.

14

I should have written this OP better. I should have started with “what are your favorite features of the safe network?”

Instead I got distracted and ended up beating dead horses.

Sorry guys, I wasn’t suggesting any delays for changes. I’ll let the horse rest in peace.

2 Likes
15

Accessibility doesn’t mean free shit for everyone. It means anyone with money or anything else of value (such as labor invested in proofreading and fixing a readme.md file) can easily obtain something (in this case, BTC).

1 Like
16

But wouldn’t that make spam attacks possible? When I proposed that all PUTs and GETs should be free while keeping the same farming reward then one argument against that was that the network would be spammed and attacked too much.

Another problem (with my model) with free PUTs is that no safecoins would be recycled and therefore impossible unless the safecoin inflation model was changed and that can’t be done because of the IPO (essentially “pre-mining” which as far as I know is a recipe for cryptocurrency failure). But the spam attack risk is a separate issue.

17

I changed the title and category to reflect the thread more accurately.

1 Like