SAFE Network - Test 15

Maybe drip PUTs rather than worry about coding a coin. And then they slow or stop the faucet if too much storage being used.

2 Likes

The only issue might/would be that the attacker invites their other accounts. Just another step in creating an account. Again most likely want something to ensure that cannot spam invites.

Or do you mean that the network (or maidsafe) provides these invites? And thus they can regulate the number and rate of issue?

2 Likes

How about account creation is removed from the tests until basic safecoin is implemented ? After all we know account creation works perfectly. Then if you need an account for testing , you request one from a regular webpage hosted by Maidsafe with a captcha and all usual spam protection. Some privilegded clients in Troon create your account and then you are sent your tester credentials by email.

3 Likes

Just got this message:

MutationFailure::MutationError::NetworkFull

What does it mean? That the network is full and we cannot add more data on it?

1 Like

Nevermind, I tried again and it worked.

1 Like

Another option would be to let the proxy_node share the hash of the IP_address that the client uses to connect. The hash of the IP could be shared with the Disjoint Group responsible. Would go something like this:

  • I try to connect to the network as a Client. The proxy_node sees my IP_address is 33.22.44.66.44.66 and hashes it.
  • The proxy_node tells tells the Disjoint Group that a client wants to connect with IP_hash: ā€œ435276c27bbbc13ad89b7f1a43bcb3c15bc43eb6313bbf2deba8fd74041aca09ā€
  • The Disjoint Group registers this data with my client_data and only allows for 1 account per IP_hash.

That way the Vaults in the Disjoint Group only allow for 1 Client Account per group while not knowing the real IP_address. The attacker could only create 1 account per IP per Disjoint Group. So with 12 groups out there he/she would fail quite fast.

This could be just a workaround until Safecoin is implemented.

6 Likes

This would help if the person is doing it all from one place, but they could be doing it distributed.

Also we should allow people to have multiple accounts in testing

Maybe allow a certain number of accounts, like 5 or 10 from the one IP address.

3 Likes

We can call that the ā€œLaw of Satoshiā€?

1 Like

Has anyone tried to run a community network node alongside an official network, do the two still conflict?
ie. can two separate network nodes run on the same lan.

1 Like

I was hoping for ā€˜the law of drehbā€™, but I suppose ā€˜the law of Satoshiā€™ has a wider appeal :stuck_out_tongue:
A cynic might say it is called human nature

2 Likes

Isnā€™t the idea of safecoin and account creation is to allow anyone with a computer to farm safecoin and thus be able to create an account? Farming is designed to be as inclusive as possible.

Yes. BUT right now someone is disrupting (or more accurately somewhat hindering) the TEST network 15 by creating a large number of accounts and filling each one to its maximum capacity thus making it difficult for others to create accounts and carry on with developing apps.
Any restrictions mentioned in the posts above are intended only to mitigate this particular threat at this time,
When the network is fully launched with SAFEcoin this kind of ā€œattackā€ becomes prohibitively expensive - as does spam and othersorts of generally ā€œundesireableā€ behaviour on the present internet.
So any proposed restrictions being discussed here will no longer be needed.

10 Likes

I agree with @Seneca and think the only real possible solution for now is a simple testcoin that has to be farmed (or just sent to you from MaidSafe etc) thatā€™s required to upload anything.

Will also let us do practice things with coins, etc. Gets us closer to multiple goals.

2 Likes

Yep.

Remember these discussions on drips/coins/etc here are for the current set of tests to try and prevent/slow down anyone from scripting account creation and fill up the network as an attack. Later on when (test) SAFEcoin the problem is greatly reduced, just have the chicken/egg problem for testSAFEcoin :slight_smile:

Hope so. But it seems that your hardware & internet link will have to be of or better than some minimum amounts. While not perfect it is reasonabe. Canā€™t expect my Intel 4004 to be able to farm.

3 Likes

The installation should be as simple as standard install using SSH on linux. That is a single command and then another to start the vault.

I am sure that a number of people will create templates for various VM providers, be it the so-called cloud VM providers, or virtualbox etc for running on your PC.

It is my hope that the ā€œminimum amountā€ be reduced from 6Mbps to perhaps, 4Mbps upload speed.

In the US, two of the biggest ISPā€™s are COMCAST[1] and CenturyLink[2]. The packages that most people buy are on the order of $65/month after your first ā€˜promotional yearā€™ runs out, if internet is not bundled with other services such as telephone. These packages have 5Mbps upload speeds. Given that you rarely actually get the full advertised speed, 4Mbps minimum upload speed would really go a long way to opening up SAFE farming to a big chunk of people in the U.S. that find the 6Mbps requirement just out of reach.

[1] Comcast Xfinity Internet Plans and Pricing Review 2023
[2] https://promotions.centurylink.com/offers/fastspeeds/

6 Likes

I am hoping for it to go down to less than 1Mbits/sec, because that then includes all the people who have reasonable ADSL2 services. There are some ADSL2 links that are not good and the line has to be drawn somewhere to enable decent download times from the SAFE network.

For the majority of Australians the maximum is 2Mbits/sec on HFC (a few ADSL2 did too) and 1Mbits/Sec on ADSL2.

4 Likes

The US is plagued by download limits though. I know some of you get unlimited connections, but from what I understand itā€™s quite rare isnā€™t it? I should imagine Europe and Asia will have a lot more farming going on tbh. It would be great to get it down to 1 or 2Mbits/s as that would make it really inclusive for almost anyone, as Neo says, many of us are stuck even lower than that for now.

The general trend is for these connections to get better though. When I spoke to my provider about my own connection (10/2) they said Iā€™d probably have to wait a year or so for the upgrades, but that they are working on it. So, even if Maidsafe canā€™t get it under 6mbit/s I donā€™t think itā€™s a big deal. In a couple of years that will be even more inclusive than it is today and it already provides access to the bulk of the effective and useful spare resources out there.

3 Likes

hi all,

been running a couple of days now, chugging along nicelyā€¦

but the log is full of this:

E 17-03-29 09:01:52.338944 Node(e975fbā€¦(111)) Missing public ID for peer a94badā€¦
E 17-03-29 09:01:52.343655 Node(e975fbā€¦(111)) Missing public ID for peer a94badā€¦
E 17-03-29 09:01:52.367201 Node(e975fbā€¦(111)) Missing public ID for peer a94badā€¦
E 17-03-29 09:01:59.609526 Node(e975fbā€¦(111)) Missing public ID for peer a94badā€¦
E 17-03-29 09:02:15.419733 Node(e975fbā€¦(111)) Have a94badā€¦ in RT, but have no entry in peer_map for it.
E 17-03-29 09:02:18.734793 Node(e975fbā€¦(111)) Missing public ID for peer a94badā€¦
E 17-03-29 09:02:18.771378 Node(e975fbā€¦(111)) Missing public ID for peer a94badā€¦
E 17-03-29 09:03:15.419936 Node(e975fbā€¦(111)) Have a94badā€¦ in RT, but have no entry in peer_map for it.
E 17-03-29 09:03:19.861305 Node(e975fbā€¦(111)) Missing public ID for peer a94badā€¦
E 17-03-29 09:03:19.930894 Node(e975fbā€¦(111)) Missing public ID for peer a94badā€¦
E 17-03-29 09:04:15.420152 Node(e975fbā€¦(111)) Have a94badā€¦ in RT, but have no entry in peer_map for it.
E 17-03-29 09:04:20.997777 Node(e975fbā€¦(111)) Missing public ID for peer a94badā€¦
E 17-03-29 09:04:21.039072 Node(e975fbā€¦(111)) Missing public ID for peer a94badā€¦
I 17-03-29 09:05:08.215337 This vault has received 9570 Client Get requests. Chunks stored: Immutable: 5152, Structured: 487, Appendable: 0. Total stored: 2871304211 bytes.
E 17-03-29 09:05:15.420277 Node(e975fbā€¦(111)) Have a94badā€¦ in RT, but have no entry in peer_map for it.
E 17-03-29 09:05:22.055049 Node(e975fbā€¦(111)) Missing public ID for peer a94badā€¦
E 17-03-29 09:05:22.089500 Node(e975fbā€¦(111)) Missing public ID for peer a94badā€¦

what is the missing public id shizzle all about?

rup

4 Likes

i assume its because this id is stuck in the routing table but is not actually contactable, should it not be cleaned up after x number of failures as part of keeping the routing table healthy?

if iā€™m a thousand miles out or talking bollox please tell meā€¦ :wink:

rup

1 Like