Hacker News new | past | comments | ask | show | jobs | submit login
Filecoin Mainnet Is Live (filecoin.io)
42 points by jmorgan 4 days ago | hide | past | favorite | 22 comments

As blockchains are slow and inefficient databases, it's only reasonable to employ them where censorship resistance is required. If no one is trying to censor what you are doing, avoid blockchains. I believe torrents have shown their censorship resistance.

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.

I agree that there are lots of problems and too much complexity in blockchain-based systems. But it's important to note that there is a real problem that Filecoin is trying to solve.

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.

I have seen this in the past year with two videos in particular. The most recent one I watched beginning to the end - it stayed on a video hosting site for weeks.. then one day I found the page saying 'taken down for hate speech' - I don't recall any kind of calls to hate - if there were, I would think they could of edited part of the vid and still left up the main gist of info.. like they do with song copyright complaints in techno mixes...

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.

Is it just censorship resistance that makes sense to put on a blockchain database? I know blockchains also have a trustless property to them. Maybe exchanging information among people who do not trust each other also has value here?

I think this is pretty cool. The basic idea is that Filecoin builds on top of the existing IPFS storage layer to add an incentive to provide storage to the global network.

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/

> 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.

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.

Thanks for the reply. And I did notice that Filecoin's committed storage was even larger than Backblaze's reported capacity, which was super impressive, even if it doesn't represent actual files yet. Congrats on the release!

oh look, an unnecessarily complex and inefficient way to store data :(

If the data you want to store is legal I agree.

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.

> start shutting down nodes

Isnt the data supposed to be resilient to downed nodes?


Do you know much about IPFS? It's top of class for distributed storage.

I'm interested to know what you think is a better value proposition?

filecoin will be a useful protcol for independent warehouses of data, all in an effort to consolidate in a pool against the biggest players. it will NOT be for the 'masses' of people, except to the extent they enjoy donating a few dollars of their electricity every month to someone's file storeage for shits and giggles.

Has anyone calculated approximate annual storage price per GB?

My best guess:

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...

[1]: https://filfox.info/en

[2]: https://www.backblaze.com/blog/open-source-data-storage-serv...

Thanks, this is helpful.

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...

[2] - https://filfox.info/en/deal/731671

Hey! Filecoin Tech Lead here, I really wouldn't take any prices you see for deals as authoritative yet. Things are gonna be pretty crazy up and down until the markets settle.

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.

> 0.25FIL per day per TB [...] in block rewards

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?

Sounds like an exciting project. What kind of data/information category do you expect people to store in the Filecoin ecosystem in the future? By information category, I mean is it medical data, internal enterprise data, politically sensitive data? Or is all data kind of encrypted so nobody has a clue as to what the data is that is being stored?

I look forward to seeing these markets settle. I really like that Filecoin actually seems to be pegged to a real-world asset, i.e. storage capabilities.

Good point about Backblaze's costs vs. consumer pricing. I was actually citing their cost as a reference for a lower-bound on how much storage can cost nowadays per TB. I wanted to compare that to the Filecoin network. However, your comparison is better anyway. Filecoin is basically operating as a network in the same capacity as B2 (i.e. as a storage service provider) and we should compare prices accordingly.

Also, good find on the filecoin deals link. That's exactly what I was looking for but never found.

> 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.

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.

I'm glad most of the network seems to be rolling their own solutions. Plus, what I said really only applies to lazy miners. It's probably not that hard to set up a thin compatibility layer between B2 and Filecoin, but as another commenter noted, they charge about $60/TB/year.

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.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact