Filecoin Mainnet Is Live
filecoin.io
filecoin.io
So if it's not for discovery or censorship resistance, what could it be for? Efficiency? Vertical integration of hard drives into the highest density 4u space is already open source. For instance 60 drives in 4u: https://www.backblaze.com/blog/open-source-data-storage-serv...
Backblaze and others also use vertical integration to get better prices from all vendors for the drives, connectivity, and they even use fancy parity systems to get redundancy across machines with minimal overhead. Distributed (horizontal) scaling loses to vertical integration.
So what are you left with? Yet another token, with a proof of storage instead of proof of hashing or proof of memory hard hashing, or proof of time locking (caveat, I founded a cryptocurrency based on that last one, for inflation only, consensus is still vanilla POW.) At least one or two "usuable" things came out of the billions wasted of 2017 ICO money... I guess.
Censorship absolutely exists in file storage. Just blocking access to a website could be considered an example of this. Torrents are one solution, but normal people don't want to set up servers to stay on for people to be able to get their files. And there's no standardized way to have someone else seed a torrent for you in exchange for money.
IPFS is basically a giant, distributed file server. Also, it uses libp2p under the hood, similar to torrents. However, it has an incentives problem. Why would other people offer you storage on their devices?
So that's the problem, and Filecoin claims to have a solution. I personally think there's too much complexity in the interface right now for people to prefer it over S3.
Anyhow - some people likely believe that some videos / news stories / files are important enough for people to have access to that they would likely chip in some coins to keep them accessible I'm sure - I would even say there is and will continue to be more of need for further distributed / multiple backups (some available in some countries and not others for example) and people will certainly be willing to pay and others will to host - to get around bans in certain countries for example.
I hope there will be a filecoin option or other options for the world to push multiple copies to multiple countries so when access is denied in one place and file removal is demanded in another - that there would be the possibility that people could access from a third place via a 4th place that kept access available.
No idea how all that would come to fruition, perhaps it's already been figured out with torrents or peertubes or similar things, I have not researched that aspect.
But I have been seeing more and more censorship in more and more places the past few years - and it appears gov's and other interested groups are ramping up for more, not less - so the need will be increasing more and more in most parts of the world I am guessing.
I also imagine there is a desire in some neighborhoods to be able to share videos and other files with the community that would be censored by news portals and national hosts - if a secure scuttlebut meets filecoin crosses with some sort of non-accessible via internet - and only shared via bluetooth / wifi / and limits on bandwidth taken kind of thing is out there I think we'll see more use and need of such a thing as well.
If you're someone who follows cryptocurrencies, the first question you're probably asking is "what's their consensus algorithm?". The answer is "proof of storage". Assuming they've figured out the technical details (I haven't read their paper thoroughly yet), this seems like a decent idea. It avoids the massive power requirements of proof of work, and there's lots of unused storage out there in the world for people to offer for a price, and they can offer it at a very low power usage. Persistent storage doesn't actively require power.
As a back-of-the-envelope calculation, if you total up the storage capacity of every iPhone in the world, you'll probably reach something like 64000 petabytes [1] [2]. Backblaze, a fairly large cloud storage provider, has a total network size of around 500 petabytes [3]. Maybe S3 is a hundred times larger? Maybe? Still, it's good that the combined storage capacity of all the iPhones in the world alone can compete with the big storage providers. That gives me hope that there's at least the theoretical potential for actual decentralization, as opposed to with proof of work, which tends to favor a minority of people with large compute rigs. Normal people don't want to max out the chips in their phones and laptops 24/7, but maybe they wouldn't mind using some passive storage.
I'm not sure about how consistently people in Filecoin need to be connected to the network, but that could prove to be an issue.
Another thing to look at would be whether there's any way for a third party to see what data you're holding. This is an issue because you probably don't want to let random other people store whatever they want on your machine. Lots of that content could be illegal, and (IANAL) I'll bet the law can hold you accountable just for storing it.
[1]: https://kommandotech.com/statistics/how-many-iphones-have-be...
[2]: https://www.counterpointresearch.com/average-storage-capacit...
[3]: https://www.backblaze.com/blog/500-petabytes-and-counting/
The initial release today focuses on Storage Mining, which requires the nodes storing the data to be online pretty much all the time to be able to submit proofs of that storage. This is because storage miners are the block producers, and we need to have high certainty that they actually have the amount of storage they claim. The requirements here will go down over time, we have a number of different proof constructions we are looking into that will improve things in nearly every dimension (zk-snark technology is still really young, and we're pushing its limits).
Filecoin (even today) has retrieval miners, that earn money for serving data as clients request it, but it needs a good deal of work still to form an efficient market. Its current state is more like Amazons 'Requester pays buckets' on s3, but is rapidly improving and many different groups are building out exciting things in this space. A retrieval miner doesnt really need to be online for any amount of time, only needs to respond to requests when it wants to get paid.
> a fairly large cloud storage provider, has a total network size of around 500 petabytes
Its worth noting that the current amount of 'committed power' (aka, proven storage) in Filecoin as of this moment is 572.8PiB, with significantly more than that waiting to get onto the chain. Most of this is just 'committed capacity', space that doesn't contain users data, but is proven to be there and ready to be filled with real data when it's demanded.
I'm interested to know what you think is a better value proposition?
If you want to store things no other provider will touch this should do the trick.
Seems like the guys who raises millions to build this did just fine.
Doubt it’s a big enough market to survive once investigators start shutting down nodes for kiddie porn.
Isnt the data supposed to be resilient to downed nodes?
trollface.jpg
Filecoin is hovering at around $65 USD / FIL [0]. The rolling 24-hr reward for storage on the Filecoin network is currently about 0.25 FIL / TB, or $16.25 / TB. Backblaze says that they manage around $36 / TB [2], and they have incredible knowledge and scale, so that leads me to believe something is wrong in this calculation I just did. Maybe the reward isn't a good measure of the price of the storage?
EDIT: I think the issue is that I didn't take into account storage duration. Not sure how to do that. Storing a blob for two years will be more expensive than for a month, and I know that the Filecoin network accounts for this. After a brief exploration of their site, it seems like you actually make bids to store your data at a certain price. I wish there were an obvious place that showed the most recent offers so people could get a good idea of how much it costs.
Also, I had a thought that a smart miner will just create a thin layer between Backblaze B2 and Filecoin to earn some FIL basically for free since no one is going to out-price B2. And then we're just back to storing everything on B2 and S3, except in a less efficient and more opaque way.
[0]: https://www.kraken.com/prices/fil-filecoin-price-chart/usd-u...
[2]: https://www.backblaze.com/blog/open-source-data-storage-serv...
The $36/TB is how much it costs backblaze to build a storage pod. Actual storage price for end consumer is $60/TB/year ($0.005/GB/month) on B2 [0], with triple redundancy[1].
It’s seems to be really hard to just get a price for the Filecoin storage, but according to some deals I found in Filecoin explorer, it comes out to about $65/TB/year with no redundancy [2].
[0] - https://www.backblaze.com/b2/cloud-storage.html
[1] - https://www.backblaze.com/blog/vault-cloud-storage-architect...
Filecoin miners right now are making about 0.25FIL per day per TB they store. Thats just in block rewards (remember, Filecoin is a proof of SpaceTime network). Any amount the user is paying is on top of that price, so basically, if the user pays enough to cover the transaction fees and bandwidth costs (very small), it makes sense for the miner to accept the deal. This effectively means the block reward is subsidizing the price of storage.
How would that change say a year from now, assuming some realistic network growth rate?
Couldn't find in your blog, docs or whitepaper, how exactly is block reward calculated and changed over time (apart from 30/70% proportion between simple and baseline mint, 200% yoy baseline growth and some other high-level overview stuff), is it documented somewhere or is source code my only friend here?
Also, good find on the filecoin deals link. That's exactly what I was looking for but never found.
They can totally do this, but they are on the hook for that data. If the backing store for the data that the miner is trusting goes down, the miner loses a pretty significant amount of collateral (currently 0.2 FIL per 32GB sector).
In practice, we're seeing most miners set up their own storage clusters. Whether a bunch of small (100TB-1PB) nfs boxes, or much bigger single filesystem CEPH clusters.
If you wanted to really be competitive on the Filecoin market, you could probably set up a rig just like Backblaze does (they open source their designs) and achieve the approx $30/TB/year mark, which is basically half of the B2 compat layer.