IPFS Definition: IPFS, the InterPlanetary File System, is a peer-to-peer protocol for storing and sharing files, in which each file is found by a content identifier (CID) derived from a hash of its data rather than by the address of a server. Because the CID changes whenever a single byte of the file changes, anyone who retrieves the file can verify it has not been altered, no matter which node delivered it.
What Is IPFS?
A normal web link tells your browser where to look: this server, this folder, this file name. If the owner moves the file, edits it or shuts down the server, the link breaks or quietly points to something new. IPFS replaces “where is it?” with “what is it?”.
Protocol Labs, founded by Juan Benet, released IPFS in 2015. In its network, a file’s address is a fingerprint of its content, produced by a cryptographic hash function. You request that fingerprint, and any computer holding a matching copy can send it to you. The data travels over a peer-to-peer network, so no single host has to stay online.
Crypto adopted IPFS because blockchains are terrible places to store large files. Writing a 1 MB image to Ethereum would cost a fortune in gas, so most projects store the file on IPFS and record only its short identifier on-chain. Understanding how that identifier is built explains both the strengths and the weak spots of this arrangement.
How Does IPFS Work?
IPFS first cuts each file into chunks, 256 KB each by default, and hashes every chunk. It then links the chunk hashes into a tree structure similar to a Merkle tree, and the hash at the top of that tree becomes the file’s CID. To find who holds the data, your node queries a distributed hash table, a shared directory spread across thousands of nodes that maps each CID to the peers storing it.
Consider a 1 MB artwork split into four chunks. The artist adds it to IPFS, gets a CID starting with “bafy”, and later edits a single corner of the image. Only the chunk containing that corner changes, so only its hash changes, and that forces a new root CID for the whole file. The other three chunks keep their old hashes, and IPFS reuses them rather than storing them twice.
That behaviour matters for anyone who owns an NFT. If a token’s metadata points to “ipfs://bafy…”, the image behind that link can never be swapped, because a different image would have a different CID. A token that points to an ordinary web address offers no such guarantee: the server owner can replace the picture tomorrow and the token will display the new one.
Storage, however, is voluntary. A node keeps a file permanently only if it pins it; unpinned data sits in a cache and gets deleted during routine cleanup. Paid pinning services and Filecoin, a blockchain that pays storage providers to prove they hold data, exist to fill that gap.
IPFS vs. HTTP
| IPFS | HTTP (the regular web) | |
|---|---|---|
| How files are found | By content (CID) | By location (server and path) |
| Who can serve a file | Any node with a copy | The server named in the link |
| Tamper evidence | Built in: altered data gets a different CID | None without extra tools |
| If the original host goes offline | File survives if any node still pins it | Link breaks |
| Persistence | Not guaranteed; requires pinning | As long as the host pays for the server |
Why Is IPFS Important for Traders?
IPFS decides whether many crypto assets keep what they promise. NFT collections, token metadata and the front ends of DeFi apps often live on IPFS, and a collection whose files nobody pins can end up as tokens pointing at nothing. Before buying an NFT, you can check its metadata link: an “ipfs://” address fixes the content, while a plain web URL leaves it under the issuer’s control.
The protocol has also shown it can route around censorship. When Turkey blocked Wikipedia in 2017, the IPFS team published a full snapshot of Turkish Wikipedia on the network. The block targeted Wikipedia’s servers, but the content was now identified by its hash and could be fetched from any node holding a copy.
Its main limitation is that immutability is not the same as permanence. A CID proves what a file is, not that anyone will keep it. Many users also reach IPFS through public gateways, ordinary web servers that fetch IPFS content on their behalf, and relying on one gateway quietly reintroduces the single point of failure IPFS was built to remove.
Key Takeaways
- IPFS is a peer-to-peer file protocol that addresses data by a hash of its content rather than by the location of a server.
- Every file gets a content identifier (CID), and any change to the data produces a different CID, which makes tampering detectable.
- Files are split into hashed chunks linked in a tree, so identical chunks are stored once and any node holding a copy can serve them.
- IPFS guarantees integrity but not persistence: data survives only while at least one node pins it, which is why paid pinning and Filecoin exist.
- For NFT buyers, an ipfs:// metadata link fixes the artwork, while an ordinary web link leaves it open to replacement by the issuer.
Is IPFS a blockchain?
No. IPFS has no blocks, no consensus and no token of its own. It is a file-sharing protocol that blockchains often use to store data that would be too expensive to keep on-chain.
Are files on IPFS private?
No. Anyone who knows a file's CID can request it from the network, so sensitive data must be encrypted before you add it. Treat anything uploaded to IPFS as public.
Can I delete a file from IPFS?
You can unpin it from your own node, but you cannot force other nodes to delete their copies. The file disappears only when every node that stored it stops pinning it and clears its cache.
What is the difference between IPFS and Filecoin?
IPFS moves and addresses files but pays no one to keep them. Filecoin adds a paid storage market on top, in which providers earn tokens for proving they still hold the data.
