31 posts were split to a new topic: Are there free things or is everything paid for?
Filecoin has passed 1 EiB in miner storage (1.034 at the time of writing).
Storj has been looking at exabyte scale for a long time, great to see filecoin showing the resources for that kinda scale are really there and motivated.
Also worth observing that burstcoin network capacity has gone from a peak of about 600 PB to currently around 300 PB. Wonder how much of that went to filecoin?
11 Likes
Currently filecoin has 1.073 EiB total storage (not the same as total uploads) across 739 miners.
For Safe Network to have this same level of consumption and node count (dubious comparison but let’s go with it) all on 1 Gbps connections, it would take 154 days fully utilized to fill all 739 miners to a total of 1.073 EiB. Not terrible, not great.
If we take an example high speed connection at AWS of 25 Gbps the network could deal with 1 EiB uploads in about 6 days.
If we instead keep 1 Gbps and expand to 10x nodes (7390 nodes) the time comes down to 15 days to upload ~1 EiB.
Some caveats:
- having a node fully saturate a 1 Gbps connection for such long periods is unlikely. So it will probably take longer than this to reach 1 EiB.
- getting 1 EiB of user data is quite a different proposition to having 1 EiB of available-but-unused storage space.
- what sort of connection speed for nodes will be common? Probably not 1 Gbps, at least not in the early stages.
I guess it will be more likely 3+ years after launch that we see 1+ EiB consumption. It’s an interesting landmark to ponder.
Thought it was interesting to look at measures of success. 1 EiB is definitely quite a nice landmark, but the context matters so I’m not sure if filecoin has given us any significant insight here. Interested to know what others think about this.
8 Likes
I see a mix of home, work, and data centres. Might be difficult to determine average speeds since there will be hops between nodes as people upload their data.
Also you need the 1EiB of data to start with and with the startup network I doubt there will be enough people to provide that.
Now if within the first 6 months the torrents people see it as an advantage to upload once for forever downloads then the 7000 nodes maybe a massive understatement and definitely more than the 1 EiB data
6 Likes
Safe network can work as low risk income for unused space in big data centres, so until they do not have better utilization they can run SN node.
Anyway it will be wise to diversify to more than one network, if you are not big supporter of Safe network.
1 Like
I claim the opposite. If you are a true supporter of the foundations behind Safe, you must support other projects. If we fail, it is important that someone else succeeds…
This doesn’t follow for me. If we all did that we would weaken the support for Safe, and atm it’s the best prospect by a long way. So by all means support other projects, but there’s no right way to achieve the goals set out in the Safe fundamentals.
I agree with the principle of diversity, and have no issue with folks putting effort into more that one project. I’m glad you and others do because it reduces the need for me to do so and I like to focus. So I focus my energy here. Also, that could change, but no sense of that at this point. It would take something very special given where we are.
6 Likes
Hey @Dimitar have you ever been able to compare or find out the fundamentals of other projects, perhaps even extrapolate them from their literature? It would be great to see this and where they all come from. I think some will be pure money earning and some will be to compete with AWS or whatever and without any prejudice, it would be good to know what works technically and also what works economy wise.
6 Likes
Support can be anything - on my list it’s to subscribe to their social channels, to like what you like, and why not run a farm on their networks.
For example, I was a farmer in Sia and provided free space there. Sia is probably the closest as a fundamentals to Safe.
But any project that teaches people that decentralization is a good thing contributes to a better future. I am currently a farmer in Noia and Storj, and although they are closer to business than to the public good which will be Safe, they also contribute by drawing attention to decentralization. 
I see the diversity in having the other projects out there to ensure all the projects hone their systems to be better.
But to deliberately have multiple “safe networks” so that at least one succeeds does not sound smart to me either. “Divided we fall” comes to mind. Data on one or multiple competing “safe” networks means more cost to upload and plenty of cases where various pieces of data only exists on a few or just 1.
If the Safe Network cannot survive then why would the others? By reducing the participating nodes in any of the “safe” networks only weakens that network. And by extrapolation all the competing “safe” networks will be weaker than just having one and getting people to all come together and combine their resources to succeed.
EDIT: obviously I am not referring to specialised “safe” networks that are created to fulfill a specific purpose.
4 Likes
That’s right, the world will be a better place if we all work together, right? But people cut off other people’s heads because they don’t believe in the same thing as them. People cannot work in large groups. We are not genetically predisposed to this. Inevitably, there will be other Safe Networks.
So which internet do you prefer? How many “Internets” are there connected to your house?
Anyhow I still stand behind “Divided we will fail”
4 Likes
I don’t think you understand. This is not our choice. This is other people’s choice. Our choice is to be as competitive as possible, and I have repeatedly pointed out what the biggest dangers to our network will be - economic vampire attacks aimed at stealing farmers. Nobody commented on that. Only I popularize this danger.
And so was the case with the early internet, intranets, MSN, etc. But only one that succeeded was the original.
The point is to make the Safe Network something people want
EDIT: and my other point was for us to deliberately have multiple safe networks as was suggested. NOT that there would not be others
3 Likes
Safe networks will be much closer to countries than to highways. Yes, highways connect all countries in Europe, but each country is self-governing…
Now, 21h later, 1.085 EiB. That’s 12 PiB added in 21 hours. 585 TiB per hour. Equivalent to plugging in a new 4 TiB drive every 25s. Eye watering.
Makes me wonder if / how Safe Network could facilitate that rate of growth. It would be amazing if we could. Think how cheap storage would be if we could get that rate of new storage coming online. And to think this is after barely a few months of filecoin mainnet being launched.
This is making me feel like we are currently aiming at a concept of a very highly coordinated on-ramp for data, and maybe we could start thinking of ways to loosen that up a bit (kinda like a layer-2 onboarding of sorts)… firmly in brainstorming territory here though, maybe even approaching the edges of what can reasonably be defined as the boundaries of Safe Network vs some other thing.
10 Likes
@mav And one of Safe Network’s issues will be the problem of knowing just how much storage is available or used in the whole network. Maybe we should target the torrent market in a subtle way with examples of how Safe Network can replace torrents.
9 Likes
I reckon statistical analysis will give us a fairly accurate answer of how much storage is available or used.
Any single vault can look at how many chunks reside within a certain xornamespace, and extrapolate from there reasonably accurately.
Spare space measurement… not so sure. I think we can still do statistical analysis based on how much the full nodes are storing and some assumptions about efficiency, eg not-full nodes will presumably run somewhere between x%-y% spare space since it’s a waste to run at 10x that amount of spare space.
When thinking of how to measure the network size and conceptualize the size of it, my mind goes to katamari (a game where you start as a small ball and you roll it around to pick up stuff lying around to grow bigger and eventually you end up rolling up whole planets). With the exponential growth of sections it seems we will start small but will become a steamroller once the section prefix is large enough, able to consume huge loads because of the massive parallelism of a big network.
Trackers are the place to pursue since they are the essential hubs that allow torrent peers to coordinate their activity. Have a look at https://opentrackr.org/
10M active torrents over the last 7 days
40M seeders
It might be possible to set up a similar site to this where the safe network acts as the tracker. I would imagine initially some kind of default-to-torrent-tracking (ie duplicate the behavior of opentrackr) but add on some option to ‘sponsor’ the torrent, ie the sponsor pays for that data to be uploaded to the safe network. Could be a nice entry point to that ecosystem?
bittorent dht is another avenue that might be worth exploring.
9 Likes
I know this project (Safe) offers more but if we talk of only storage I am afraid of ‘first mover advantage’ becoming bigger and bigger there.
My concern is even if something 10 times better than Amazon is developed now it is unlikely to see the light of day.
Could their success overshadow what happens here when it goes live…
Am I unnecessarily worried?
2 Likes
In my opinion, the struggle is who will be the first to take the free resource from people’s computers. Whoever is first will have a huge advantage - that’s why our direct competitors are things that can be farmed with ordinary computers - Sia, Storj, Noia.
The problem with them is that it is not enough to click one .exe and receive free money from your free computer resource. To get them to work requires serious (for the average person) knowledge…
2 Likes