My pronouns are Aye, awright and whatever.
3 Likes
Absolutely.
For a well functioning market, I expect it would be advantageous to have some kind of decentralised signal that incentivises nodes to join in a specific part of the network that needs capacity vs an area that doesn’t (e.g. if nodes in a group need more nodes in their region to be competitive / gain capacity, they could offer a bounty to encourage joining nodes to choose their area, and joining nodes can either join the first address space they are given, or try a few more times to ‘bounty hunt’ to join an address space of significant demand… existing nodes could also scout for bounties to decide whether to re-start & re-locate to an area of the network they’re more needed).
Another function that would benefit the market / even out node fullness would be if uploaders had some choice to avoid high-fee areas (E.g something along the lines of; self-encrypt a file twice, once with a nonce, then compare quotes for each chunk, upload only the cheapest of each chunk, and then add notes to the data nap to ensure downlaoders get the full file correctly).
I’m loving the progress with the test nets etc, but, rightly or wrongly, I feel the lack of attention to market dynamics is the biggest risk for Autonomi as it launches… as technically & marketing wise, the team is killing it 
1 Like
Pretty good estimation 
Sorry @Josh - the shu will most certainly come with larger waves than that…
2 Likes
After seeing the speed they came on at I thought id fire up some nodes with out an interval on start up
in vdash it doesn’t look like it went that well lol
nice to see a large increase in the network size 
4 Likes
Maybe but so far it appears to have been around 5000, 2000 maidsafe, Neil normally has around the same so a 1000 odd among others.
The only other person (we know of bringing larger numbers) has as yet not made testing waves.
3 Likes
I hit start and left our flat … Need to check if everything went well
They joined fast but did resource checks all the way and just added nodes when resources looked good
1 Like
Let us know the results my no interval test appeared to be a shunned mess. I’m unsure if the nodes would keep going and earn after a good shunning 
Its you? 
dammit why you all going to make me work so hard to remain competitive.
5 Likes
Lol, I will come for sure… I didn’t upgrade my home internet for it to be not used… but not today! 
.
Been a busy week on all fronts, and I suspect it will continue to be busy for the foreseeable future… 
.
10 Likes
I’m looking forward to seeing that network graphs upward trajectory when you start getting down to business
5 Likes
hmmm - bit of a mixed picture at my side - shunning - yes - but not as extreme as yours …
…the most significant lesson so far: unbelievable the differences in resources of VPSes … some can just stand 12 (!!!) nodes … others are okay with 100+ … (same hoster - same configuration …)
…and not sure the earned nanos number of 0 there can be right … i need more time to have close enough looks at that stuff -.-" … this beta launch is very badly synchronized with my life …
3 Likes
I know right, so many years of when is this going to happen straight to slow the f down please. 
6 Likes
yes same results here vps’s are vey hit and miss iv got ones that can run 70 no problem and others that get upset at 30 wit same config and provider.
i would be more upset about it but the network is not designed to run on vps’s in a bit barn so maybe its a good thing for the network even if not for our node counts lol
2 Likes
within that timeframe network size increased significantly … but the spots of the network with high prices seem to remain (ignore the dip - that didn’t happen … just an error in the estimation process i guess)
…but then again … i can only get the current price out of the log files and that just writes it when someone gets a quote for uploading new data … so reliability of that storecost graph is very questionable … would be cool to get the storecost from the rpc api
3 Likes
What are your vps specs?
I would think Maidsafe have this worked out pretty well.
That setup costs $1200 a month to run.
40 nodes per droplet.
1 Like
Those are interesting ideas and may be deserving of some exploration. It could be that adding in a way to officially game the system would lead to it being abusively gamed though. But maybe gaming is needed in some way. What would an ant do? Their actions aren’t random because they signal to each other about where the resources are.
1 Like
most are 10gb ram 4 vcp 150gb ssd but its not digital ocean so who knows what’s going on at the provider iI with
1 Like
Yes, it’d be important to ensure gaming isn’t beneficial, but that responding to incentives is.
With the ant colony analogy, there needs to be pheromones to tell other ants where to come to feed/work, but at the moment it’s random.
Previous discussions on this have ended with ‘hopefully it’ll all be fine at scale’, but I know that I’d tend to distrust that things which fail on a small scale will sort themselves out at scale… I’d find it easier to trust that it will work at scale if I saw an efficient market operating well on a small scale / shrinking network situation.
The examples I gave are rough, but I expect there will be ways of incentivising nodes to join ‘busy’ areas of the network, and uploaders to have some choice between competing nodes, without making the system more vulnerable.
2 Likes
That’s the nature of Cloud servers. One VM can end up on a server, switch or set of disks that is being flogged like a rented Transit and while another is on a relatively lightly loaded set of infrastructure. If you are running an undemanding service you will never notice but if you are demanding a lot of the resource that is constrained you might and it might impact your service.
3 Likes
As in “awright time for nother round”. and “whatevr” in response to “your turn”
And “Aye” in response to “want nother”
7 Likes