Hey, I built a private network with IPFs (version: Release v0.4.22). I used “ipfs bootstrap RM all” to delete the default IPFs nodes, and added the key created in the ~/.ipfs directory. Then the other nodes were installed in the same way and connected to each other through “ipfs bootstrap add” to form 10 IPFs nodes. Private networks. But when I uploaded files on one of the nodes, I found that the files were not segmented and stored on other nodes. Even other nodes can download through HASH values. What’s the reason?Does not the file uploaded by the IPFs node be segmented and stored?
Welcome to the forums!
That’s not how go-ipfs is designed to work. IPFS-cluster might support doing something like this in the future, but it doesn’t currently support scattering and distributing chunks for a single pin across peers in a cluster.
Did you read something somewhere that gave you the impression that this is how go-ipfs currently works?
Sorry, maybe my understanding of ipfs is wrong, that is:
- IPFS does not support distributed storage. For uploaded files, it will be stored locally and generate HASH values. Can other nodes retrieve and access these data through HASH values?
- What does the so-called slicing function mean, instead of storing data in multiple nodes? If not, what is the advantage of comparing cloud storage? Isn’t the main purpose of ipfs not to store large data?
- So the main function of IPFS is to encrypt files to generate HASH values, and to find files through HASH values, as well as to achieve permanent storage of data, or even replace HTTP?
I think you might be interested in following ipfs-cluster. DAG sharding is one of the things on their roadmap.
If you’re only using go-ipfs by itself, then this is correct.
Can you give more detail on what you’re asking about in the list of questions in your second point? Are you asking about why data is chunked?
Your first point isn’t accurate. go-ipfs doesn’t encrypt your files for you; but you can encrypt them yourself if you want.
Some of IPFS’ high-level "why"s are documented on this page: https://ipfs.io/#why. One of the main benefits (in my mind) that is missing from your list is p2p-based distribution of content. If you already know the multihash for data you want, you don’t have to rely on the original uploader to provide that content (and at the same location you’re looking for it). Rather, you can download it from one or more peers who also have it.
Ipfs doesn’t store files.
It use case is to reduce bandwidth cost on centralized storage.
If a 200mb file is viewed by 10 people on the cloud, 2g pricing (0.2×10) bandwidth is to be paid by the host to cloud provider “costly”. But ipfs solve the pricing problem by encrypting the file, then the user can retrieve the file through the hash.