Hello,
you guys talking to much about what icp can do, but then why you not creating simple cloud storage,
where we can safe files, video and pictures
IC drive is ghost town
is there any cloud storage that is working?
5 Likes
You can do all of this within an ICP canister smart contracts. You learn more on how to build them and the capabilities here.
I don’t believe ICStorage is active but:
4 Likes
https://nazwh-uiaaa-aaaai-qbexq-cai.ic0.app
like u wish to use!
What is the usefull thing on icp? or is all just like the same story
Why are you waiting on DFINITY to build this? This is application level stuff. Just build it yourself Or ask community developers to build it, but you have to convince them that there is a demand for this.
2 Likes
It’s a cloud platform, that’s one way I like go describe it. So if you’re a developer you can develop on it. Storage is right there for you in your canister/application through a simple API.
7 Likes
This is a common communication gap that I struggled with. I think cloud compute is where ICP shines, whereas storage is something it can do as a side effect. It’s not what it’s best at, though people have worked on this with ic-drive, w3disk. Juno also is a DB as a service for text and file storage.
But for heavy storage, other places might be better. I use ArWeave for all storage and ICP for all compute (trying to get this to catch on as the AI stack).
3 Likes
can someone build heavy cloud storage on icp and if not why?
Why dont we just build storage centric subnets instead?
Not sure why this should be subnet specific, just to clarify increased storage is on the roadmap: Roadmap | Internet Computer
This combined with e.g. an Asset canister offers large file storage similar to Arweave.
But as of now, storage on the IC isn’t immutable static storage like Arweave but dynamic data storage that allows for fast compute of the stored data on chain at the cost of higher storage costs.
For long term affordable immutable file storage similar to Arweave you’d need to either simply integrate with Arweave (already possible with http outcalls) or implement something similar on the IC.
Coming back to the original topic post, I’d love to see a dapp on the IC for file storage. Once VetKeys becomes available, it would be possible to use IBE to encrypt files securely before storing them either on the IC itself, Arweave or other.
The main strength of the IC here would be in making this decentralized file storage application available, where the files are actually stored doesn’t really matter that much anymore if they’re securely encrypted.
More information on VetKeys and their use case (file storage is an example mentioned here) can be found here: vetKeys | Internet Computer
Edit:
Looking at the future roadmap items, “blob storage” seems to be listed there, which should allow for more affordable file storage compared to current costs.
5 Likes
wait why should man use arweave, the whole promise of icp is to build on the network and not really on third party
In the long term probably yes, but in the short term they’re two different products with two different use cases from a developer perspective.
Integration with other chains and other services makes the IC more applicable for developers and ties it into other eco systems. For example seeing the IC brings smart contracts to Bitcoin is a recent interesting development.
I don’t think the main challenge is file storage by itself , the main challenge in my opinion is to build a user friendly cloud storage that takes advantage of the unique characteristics and features the IC provides.
1 Like
but then its very misleading, when all they doing is the same think they are criticizing, why should they not wait til they bring those application, all they doing now is really on a third party but also they providing support to them
I do not recommend using a blockchain for public or private media file storage. Decentralised/distributed storage is already solved just fine even without blockchains. There needs to be feature-wise innovation to justify the overhead in UX complexity that comes with building on the blockchain. The IC is also unique in its billing model insofar that it assumes the deployer of a service also finances its operations. On the IC, the hosting infrastructure is bound to coincide with the subnet / validator infrastructure. Which is a sub-par fit for something like a massive storage cloud. The IC is not suited for providing competitive storage prices, or competitive hosting scalability. Having something on the blockchain requires that there is someone paying for it in crypto. In the case of the IC, the naturally intended model is that the developer refills the cycles on the canister to keep its services available. Or you create a personal instance of a storage canister per user, and then carry the costs yourself. But that has bad user experience and requires a user to have ICP tokens before even starting to use the service. Which means doing KYC on an exchange, depositing fiat, buying ICP tokens, setting up a wallet, deploying your own canister. Nobody wants to go through these steps just to store a file somewhere. The IC technically cannot provide you with competitive storage prices, scalability, nor with competitive user experience compared to non-blockchain solutions. Neither can any other blockchain.
Blockchains are made for business interactions between mutually distrusting parties, and all such use cases require provability of all wrongdoing. You cannot prove that someone is refusing to serve you data. You cannot prove that someone wrongfully accused you of refusing to serve data. You cannot even force someone to publish your encrypted data, because you cannot prove you sent that data to him in the first place, unless you did so on-chain. Which makes the IC the bottleneck for uploads and downloads. For any widespread adoption, that is not possible. You cannot build a reputation system on a blockchain where having many identities is trivial. All these facts make a cloud on the blockchain unable to compete with conventional services. As soon as you introduce a trust/honesty assumption into your platform design, using the blockchain is no longer your best option. So either you have an impractical trustless cloud, or you have a trusted cloud that is no longer a good match for what blockchains can offer.
The only storage that is useful on the blockchain is something like DNS where you have changing records and it’s important to have high availability and integrity and liveness guarantees, and need access control for updating the entries. All other kinds of storage usually have some inherent pre-existing incentive/interest structures already in place, and a bespoke solution for an individual scenario would be much more useful, and would allow for unique features that leverage the unique incentive structure that is already in place for that scenario. For example, you probably have an interest of storing a copy of your private data on a small home server anyway, and most likely just want improved redundancy and availability via the cloud service. Or for public data, there usually are people other than the author in whose inherent interest it is to make sure that data continues to exist, even if the author does no pay them to do so, or decides to delete the files from the cloud storage. Just these two properties alone already lead to a wide range of features that an on-chain storage solution cannot leverage.
I can understand why blockchain-based cloud storage is not widely used and why there aren’t lots of people building one.
So something like Arweave is not viable either?
I just had a cursory glance at it. It seemingly allows you to “permanently” store 100MiB of data for each account you make, as well as 1GiB of modifiable cloud storage. It looks like you just need a unique email address and that allows you to make an account. If you have your own domain, you can have limitless email addresses. You can force the system to “permanently” store lots of garbage that way, for free. Public goods aren’t a right, they are an obligation. If you can force “the public” to do something for you, without any obligation for you to do something in return for “the public”, then the system is unbalanced and cannot sustain itself. “Permanent storage” means infinite duration of storage, and that it is entirely unable to ever forget anything. A system that can only ever grow, is unsustainable. Almost nothing is worth storing “forever”. It would only take a little while to grow the entire system by a terrabyte of “permanent” storage if you automate the process. Hardware costs money. Anything that the public deems unworthy of being maintained, should decay.
The idea of a permaweb is missing the point. Instead of forcing “the public” to do something, it is much healthier to give everyone the ability to voluntarily provide specific public services, such as storing/replicating specific data. And someone who contributes nothing also can’t expect to receive any services. So either you contribute valuable data which others are interested in storing, or you pay them to store it for a while. But all this is temporary. As soon as nobody is interested in storing that data anymore, and/or as soon as you stop paying anyone to store it, the data should be rightfully lost. In 100 years, nobody wants to store petabytes of garbage.
There is about 5,530,543,487,713.28 bytes stored on-chain with ICP
Some of that is files, some of that is video, and some of that is pictures
+5 Tebibyte -
This is by far more than any other known public blockchain -
I think we can close this thread as SOLVED 
5 Likes
On all this text you didn’t even mention the value property of storing on chain, which is??? SECURITY, AVAILABILITY, TAMPER PROOF.
Of course some people don’t care, but never forget the world is evolving and cyberattacks are increasing, once apple icloud or AWS, google cloud, get their private keys of the encrypted files compromised, it’s done, everyone will be migrating to new services. There’s no scalability issue on the storage side on the IC, or someone from dfinity can correct me If am wrong.
1 Like
Tamperproofness is also achieved in torrents. Availability is also given for torrents, as long as anyone at all (1:n) is seeding it. It’s not a blockchain-specific feature. And “security” can mean anything and everything. Are you referring to Confidentiality? Access control?
And regarding scalability, we are talking low order terrabytes right now. What about petabytes? or an exabyte? That’s not what the IC was made for, as far as I understand it. Even just storing all games that are on steam would easily dwarf the entirety of what is currently on the IC. And that is legitimate data you’d want to store in a cloud. Now what about the entirety of youtube? Twitch? Archive.org? Twitter?
1 Like
BitTorrent is a decentralized file-sharing protocol, but it lacks inherent tamperproof mechanisms. Files shared over BitTorrent are broken into small pieces and distributed across multiple peers (users). While this distribution method ensures redundancy and availability, it doesn’t provide guarantees that the data hasn’t been altered.
BitTorrent uses SHA-1 hashes to verify the integrity of each piece of a file during the download process. When a user downloads a file, the client software checks the hash of each piece against the original hash provided in the torrent file to ensure it hasn’t been altered during transmission. However, if the original file itself was tampered with before the torrent was created, the BitTorrent network won’t detect this.
Since the files are not stored on a blockchain or immutable ledger, there’s no protection against someone modifying the content and re-sharing it. In other words, while BitTorrent can verify the integrity of downloaded files against the original torrent, it cannot ensure that the original file itself hasn’t been tampered with.
The security of files on BitTorrent largely depends on the trustworthiness of the seeders (those who upload the files). If a malicious actor seeds a tampered file, others who download it could unknowingly receive altered content.
ICP can scale just like the internet or the bit torrent network, but perhaps you have to consider what data needs to go where because admittandly the high quality product is reflected in a higher than normal storage cost.
For example…
Why would you want to store the entirety of youtube when this data is mostly junk and is stored with redundancy by Youtube (Alphabet). I do not think doing so justify the cost.
However developing a new application 100% on-chain for video distribution can be justified due to the value on-chain technology creates for the content creators, where in theory the users can share billions in revenue created by the platform, rather than having one mega entity make the majority of those funds while creating really poor user experiences and forcing policies on their users no one needs or wants.
Also consider availability… torrents I search for often takes minutes, days or even weeks to load. While on ICP everything loads immediately at high speed because its replicated on several high-performance web servers in high-performance data centres.
Obviously… a torrent application can not host a decentralised online application, so if you require this use case (potentially worth many trillions) you can not go to torrents.
In this sense RmbRT, yes, just as you anyone of us loitering on the Dfinity forum, yes, you correctly did find a valuable use case that is under appreciated with a lot of potential. Just because you are early and thousands of people haven’t gone before you does not mean the network is worthless, it means you were here early.
But yeah… dont replicate youtube on ICP , and dont share the entire youtube on a torrent network, its not a great use of your time in my opinion.
1 Like