Ensuring SAFE's Economic Viability: GFW Edition

So, ladies and gents, I’d like to throw out a bunch of

confirmed facts about cryptocurrency & China:

  • Because of China’s currency controls coupled with its:
    • Size
    • Entrepreneurial population
    • Relative technical expertise
    • Desire for “sound” non-manipulated money
    • Population accustomed to breakneck innovation
    • Relative Affluence

China is bitcoin demand, globally.

How much of bitcoin demand, you ask?

90%, the last time I read anything on the topic. So if 10 people are buying coin, that means 9 of them are tech-savvy Chinese looking to make financial bets otherwise prohibited. I think that’s a pretty big deal.

Next, because of its internet policies:

  • VPNs are a billion-dollar industry
  • An endless cat and mous game is played between users and censors
  • Censors have been winning for a little while, though lantern is quite effective generally, it is getting less so.
  • But looking at this here application, MaidSAFE, I see the potential for a coin-paid global delivery optimization and circumvention network.

If I am groking correctly the way that the nodes get in touch with one another (some similarities to consul, no?) then this could even be run as a separate application that interacts with MaidSAFE. Participants could be paid a market rate for BW and compute, directly in coin. I do not know if this could be accomplished without compromising in some ways anonymity of at minimum the nodes providing the VPN. But then again, seems we could have plenty of intermediaries. Anyway, my thought is that this would naturally create a truly gorgeous autonomous org.

an autonomus org, that never rests, fighting for humans’ network freedom

/me offically soliciting community feedback on how to implement-- with or without using SAFE. Oh, and if you’re wondering why this is such a big deal to me, just imagine the internet without google. And facebook, and any of their spunky freedom-oriented alternatives.


I’m aware that there’s some prior art here in the form of VPNCoin, but as a user (on both sides of the market, in fact) I can say with some certainty that VPNCoin wasn’t really as sophisticated of a solution as is needed. It suffered from all of the problems that VPNs have in hostile network environments, and more. Essentially neither th economic model nor the technical service being provided were advanced enough to make VPNCoin truly useful.

What this looks like in my mind

I’m thinking of this as something that could be run with a golang or rust binary and a config file that uses a bockchain or something like a blockchain and the very latest in VPN-age, including its own protocols, and protocols for making protocols programatically, and active exploitation of “smuggling” routes that become available (eg: if you are behind the golden shield, and so am I, but I have far more BW available, then I should smuggle BW to your node in exchange for small payment from your node).

On the server side, it shoul dbe dead-easy to start new nodes and sell their BW and Compute time. Prices should be set fairly high-- specifically, they need to be high enough to truly incentivize small and medium hosting providers to begin taking this network seriously as one of their income streams. I happen to know that for most hosting companies this exact pattern is alraedy a large segment of business that gets deeply under-served.


  • Helping China solve its Internet problems could be a way to massively boost the economic viability of SAFEcoin.
  • China IS global demand for cryptocurrency. The rest of us are just 10% of it.
  • China exports demand for VPNs in the form of its support for GFW-Pakistan and GFW-Iran, and others, too.
  • Maidsafe mgiht be able to use tis network to earn directly from people in China by selling VPN service.
  • This might expose individual SAFE nodes, so it might be best to keep this and that separate.

Implementation Details

  • Multiple existing VPN Protocols used
  • Novel Protocols generated and purchased by the autonomus org automatically. Payouts based on success rate after some inital lump sum.
  • Multiple existing encrypted proxies used
  • Multipathing used
  • “Fast Datapath” used (check out weave.works and wha tthey refer to when they talk of fast datapaths)
  • A wallet setup for the client side that accepts *coin and allows th euser to choose what unfettered access is worth to them at any given time.
    • Take this from someone who’s experienced multiple situations where this is surely the case:
      • At the right time, there’s no debate as to weather Internet access is worth in excess of $500/hr. If not having it can cost you $100,000 in an hour, then reasonably, getting it for that hour might be worth as much as $90,000.
      • This would incentivize those providing access to be quite innovative in their delivery attempts, because access would be priced suitably high.
    • High Prices are somewhat desirable because they will feed R&D.
  • The Majority of the time, access would be far less than $90k/hr. But every single day that passes, I bet >90k/hr is spent on circumvention tools in China. Fact is that I’d rather our community frolic in that value stream than exist just outside of it.

Do you have more detail for this that you might give? Is the appliance that you mention elsewhere a part of it?

I gather you are proposing a VPN service for China using SAFE to traffic the data across the GFW of China?

So people behind the GFW of China would run your APP which gives them the equivalent of a VPN on the normal internet without the ability of it being blocked?

If things go as hoped by many SAFE will eventually (after years) become the new WEB and your VPN APP for the most part become unnecessary.

OR is there much more that I missed in my read. (I saw the parts about coins)


I don’t understand it correctly. Why are u proposing existing VPN protocols etc while you have safe in between ?

The way I see it is you outside the gfw run a safevpn app locally which is a proxy between safe launcher and your internet. You send the Chinese request to the net, proxy catches answer, sends it to your launcher, then gets broken up and sended to our Chinese friend who’s launcher compiles it and sends it to the browser.

Am I somewhat correct or am I making mistakes…?
Only speed losses specially ajax… how about security of the exit because people will start exits to compromise passwords etc, ideal mitm ??

Although I like the idea…

… I think I get what you mean … you always run the exit on the server hosting the website itself aren’t you? So then if done correctly everything could stay encrypted even on the webserver local host 2 mysql etc … you could implement eventually a ‘safe compatible’ logo on the site and you let the exit publish it’s metadata to a nosql db on the safe network where you can create a search index ?

Yeah, I guess this is a pretty dense concept, because 2 people are asking for more info, so I’ll do as best I can.

One of the things that a setup like SAFE can be used for is called service discovery. The nodes need to be able to talk to one another and exchange information. Now these days one of the key problems with avoiding GFW is the fact that it can identify tunnels, vpns, etc and eliminate them over time. If, however, one was to combine SAFE with several of the latest VPNS, you should theoretically be able to jump between them-- VPNs and SDN tools alike in fact.

Then, by jumping between various VPNs and SDN and proxy systems, you should be able to achieve a picture of aggregate performance. Each node should self-optimize and it could gossip with other nodes in the same role (client or server) about what is working best. Lastly, when one node is much faster than another, it could start a “horizontal” connection (if we assume that everything that crosses GFW is “veritcal”) between two nodes that are in the GFW-zone. It could then share the pipeline.

Now, there are many technical reasons why all of this is very difficult, not least of which are SAFE’s protections for nodes. Basically, I’d like to get feedback on the design, and publicize the design, so that even if I never make it, the idea is out there for anyone to use.


Looks good. I have established feet on the ground in China. How can I help?

1 Like

A search term to start with is QoS (Quality of Service), read up on it and then think by analogy.

QoS is a function of business-class routers such as those from Draytek and Cisco, and I’ve set up a few. It is filtering down to even domestic routers.

I’ve now learned that the same term refers to allocation of disk resources in modern raid arrays/disk farms, which i know almost nothing about. It appears to be an area of ongoing development.

I found a simple introduction to QoS for the home user, that gives one a basic orientation.

EDIT: Continuing the brainstorming: Crust and NAT traversal do part of what you describe, opportunistic getting from A to B (note to self: read the relevant RFCs) but without regard to cost/importance (as I understand it). So the ideal would be a combination of that with QoS’s apportionment of bandwidth resources.

EDIT1: More brainstorming (making it clear in case anyone objects to such ramblings): There are several ways to allocate scarce resources, and some examples are:

  1. Top-down optimization such as linear programming. I actually did a couple of units of this, with good marks. It is much loved by commie planners since the 40s.

  2. Top-down optimization by feedback, a field known as control theory, which I also did a unit of with good marks (so many useless things I’ve learned, lol). Cybernetics is a related term. The Chilean government under Allende was really into this “this”).

  3. Some form of exchange, which serves to match up the scarce resources (get as much for them as possible) with the cheapskate buyers (pay as little as possible). Exchanges exist in such types of resources as “commodities”, even electricity across continental regions such as North America, and it is a much studied field. But I only have a lay understanding of it. I do feel that it could be applied to the problem that you describe.


I, for one strongly encourage productive ramblings.

1 Like