Guaranteed to get right up the nose of all the Brexiters <–this is a GoodThing but I’d rather not have a Butchers Apron flag of the colonists displayed on my screen at teatime or any other time
2 Likes
This may well be true. I certainly hope so…
BUT… most of that testing will be done from DO droplets or other datacentres. This is like testing a new car only on the smoothest of motorways in perfect weather, not in city traffic or on cobbles or ice, desert etc etc.
3 Likes
thats 6 nodes out of a claimed total of 65-70k (according to my Formaciao)
As you can see I have 45 from home with a dedicated 2TB drive and a pretty decent router.
I bet the top 20 in the latest leaderboards are still all running out of datacentres. We are
only fooling ourselves - and lying to partners if we think this is OK
4 Likes
I agree with many points you say
And all this testing and results are just theory as long as @happybeing can’t use the network via mobile Internet (secure access for everyone… With a wired Internet connection =D ideally fibre)
Buuuuut I think it has been shown that the data network is indeed pretty resilient… If we now get better incentives than the temporary leaderboard stuff, and optimize bandwidth usage further +make sure people can really use this via mobile Internet too… We’re on a pretty good path… Looks to me significantly more promising than the other ‘decentralised’ stuff I’ve seen within the last years…
14 Likes
Broken promises without a mention happening again and again. Like “We get the whitepaper, API, hackathon, marketing, partner announcement, wadeva… next week” And when they don’t, there’s no acknowledgement why they didn’t. Then at some point some of us go mad, and they react with admitting poor communications. Then that poor communication goes on. Then someone gets mad…
You are right, and I’m running with commodity hardware too. But I suspect that the up/downloading, especially the downloading is nowhere near the amounts we would see in the actual network.
More centralized than ever, too.
8 Likes
Well, they can’t choose which nodes are being run and on what hardware. Other than maidsafe nodes, of course.
Maybe the rewards have encouraged data centres. Maybe this will happen after launch too. It should be cheaper to run home nodes, but maybe the demand will require more muscle at first.
Either way, we’re getting to see how the network operates with nearly 100k nodes. That’s a pretty epic test in itself, considering routing and rebalancing that is constantly going on.
9 Likes
And I in turn agree with much of what just said.
As long as we don’t work it too hard. It might work well at scale but we have zero hard evidence for that.
As I said above, so far this network has only been tickled under the chin when we should be giving it a proper kicking and trying our utmost to find ways of breaking it. Cos if we don’t, someone else certainly will.
8 Likes
=) oh well… But I don’t think we’ve ever seen data loss (not even on the crashing network before we started babying it)… But then again the giga nodes have never been full and we probably had insane amounts of chunk copy’s laying around =D Blockchain style xD
4 Likes
Oh and tbh I don’t see an issue with data center nodes being a thing… Commodity hardware and existing Internet connections are unbeatable in price (cost effectiveness on earning) but when the price rises for data center nodes to be profitable it just means they’re necessarily atm to provide enough space at a good enough connection 
… Data center nodes will vanish automatically when enough home nodes are running (at significantly lower costs) for the network to survive on home nodes only…pretty basic market mechanics at work… And since home nodes do run at a significant cost benefit (I don’t think there is doubt about this part…?) the network and its envisioned payment scheme does neither favor large farms nor data center structures as we know them… So the spreading out part should happen by itself in the long run…
… Short term data centers and servers help to provide the needed storage space +connection speeds…
6 Likes
This is not true. I don’t use datacenters to run my node, everything is from home when openly participating in the live network, and yes I am part of that top 20 on the leaderboard.
12 Likes
Cool. What kind of router and connection?
1 Like
Apologies
perhaps “The vast majority of the top 20 in the latest leaderboards are still all running out of datacentres.” would be more accurate?
6 Likes
Perhaps, and perhaps not.
4 Likes
and completely irrelevant 
… home nodes will take over in the end …
4 Likes
thanks for everyones efforts at autonomi and forum for a great read all day ever day 
A pretty resilient network is a big win, just keep iterating on ,one day at a time, every day that passes the timing and opportunities for the software and everyone involved/community at whatever level get greater ever day not less so it’s all good 
imo
5 Likes
Cant disagree with that — but it hasn’t been tested. @neo posted the other night a detailed critique (which I will link to when I find again) about why a network running with a majority of datacentre nodes would behave quite differently from one with say 10-15% of datacentre nodes.
6 Likes
I’m absolutely sure this is true … but with the financial incentive clearly favors home nodes this will result in a force towards larger home node participation … and therefore optimizations that will help with overcoming issues that come with higher home node percentage … we should always optimize the networking stuff to run really well for home networks … but datacenters can help overcome the current lack of this and help jumpstarting this imho … (if everyone would run on --home-network then we wouldn’t have relay peers … port-forwarded nodes and datacenter nodes do help other nodes to connect and do the heavy lifting currently …)
2 Likes
You mean this?
4 Likes
No - it was another post…
But anything from @neo is worth reading, often 4 or 5 times so us normal humans can understand it.
9 Likes
Fast enough, more details can be found in Advanced Router forum topic (previously discussed).
5000/5000 Mbps, 200TB Storage Allocated, 100+ CPUs, 1M+ Active Nat Session Tables (above is 1 location out of several home bases).
Under utilized hardware at the moment due to power constraints, 
.
8 Likes