I read in the FAQs post that “We love IPFS because it’s not a blockchain”. For someone who is a complete n00b regarding both IPFS AND blockchain, could you please elaborate on this? Or maybe point to somewhere that does. With all the hype about blockchain being the next best thing (as a n00b as I said), I found it striking and am curious.
What a tremendous question, @miren — I’m sorry for the delay in getting a response to you. This ended up being a much more expansive answer than I’d first planned. So, here goes.
IPFS is an upgrade to HTTP, which is the protocol your web browser used to bring you here. When you type a website into your browser, it “asks” another computer – usually, but not always, your internet provider – how to get to the site, and then loads the page.
This can be problematic, as it centralizes how we access information to a handful of companies.
IPFS removes the need for that middleman, as the request is sent to the entire network. This network is made of other users like you. When you “ask” for a page/object, the network point you to the correct place, no centrality needed.
What makes IPFS incredibly compelling for Anytype is how it gives an address to every unique object. If you’re looking for an image, for example, you won’t need to request an entire website – with all the nonsense found these days – just to view an image. You can request that object and the network will point you to it. This enables Anytype users to build a personal web of knowledge, with the ability to share and connect with others on their terms.
When we talk about blockchains, we’re referring to something like Bitcoin or Ethereum Classic. In these cases, a blockchain is a massive file that contains a record of every action that’s happened on the network. Every user of the network keeps a copy of this record. This record cannot be changed. In IPFS each network node stores only content it is interested in, plus some indexing information that helps figure out which node is storing what, so you don’t need to copy the whole internet.
I have some questions on IPFS too- i wonder if this ‘from user to user’ and ‘no middleman’ is actually true, in a literal sense, or a simplification. Can someone clarify?
As far as i understood/saw the internet, any connection to a website have to go through hoops and loops- routers or servers reroute ones request until it reaches the server where the data actually is.
But i had the notion ones own isp/internet provider was a inescapable route, by the very nature and way the internet is provided to us; Such that even vpns nor tor networks, as far as i know, cant ‘not pass through the isp’- all they do is sort of masking what is passing through and to where, with encrypting and using different routes… in vpn from you to isp to a proxy/server somewhere then the site(so the destination sees you as coming from somewhere else), and in tor from you to the isp to the first node (or whatever is the correct term)- and from there on who knows how many more nodes.
Or am i misaken? Can packets go straigh from user A to B without the isp in between?
Because if not then, at the end of the day, theres would still be a middleman watching over wouldnt it?
In another framing, what i mean is we dont have cables connecting User A and B, wi-fi is short ranged so the only noncable venues depend on enterprise level networ and providers- so the data have to phisically follow routes that pass through different servers (and i assume our isp)… so im not seeing how these claims about IPFS can be taken literally- and since most information i can find (shortened, without diving into the details) shortens it like that i wonder if im missing something.
If that is the case then isps would still see youre asking CID/ipfs id x(content x), and from where the packages are coming from; So its encryption? But data can also be encrypted over http doesnt it?
Another claim i dont quite get is when people say in IPFS ‘content cant be taken down’- but at the same time it claims that if im hosting content and broadscating just the id of it, the peers dont hold my content- so if i go down or delete it it would be down wouldnt it? the ids would point to a host that no longer holds that data anymore… it would only still exist if someone somewhere made a copy of it (that, being hosted via ipfs/id, would still be reacheable by the same id i guess). Correct? So the ipfs wouldnt be spreading the web (its data), not by itself, bu only if people were spreading the hosting of said data as well…
im still confused at grasping ipfs, even if i understood the whole id/dht part.
The simplest way to understand IPFS would be that 10 friends sit around each other in a classroom, and one of them requires a pencil. If this were a traditional web 2.0 implementation, the request would be sent to a friend who has a box of pencils; you might not require the entire box, but the request will still be sent to them; this process is time-consuming and resource-intensive. On the other hand, what an IPFS implementation would do is send the request for a pencil (similar to shouting in the class that could anyone give me a pencil) and the person nearest to you would give you the pencil reducing request time and increasing speed.
The notion that IPFS is ‘from user to user’ and ‘no middleman’ is not completely true but is not false either. The watered-down explanation would be that there will be nodes that act as middle-man to transfer data, but they do not act as the single controlling entity through which all data requests have to be passed.
The way IPFS works is that there are nodes that know other nodes on the network. So they are connected like a mesh.
What an ISP does is, forwards your data request to the DNS and based on your request/header, there is one central server that responds to your request this means whatever you request is visible to the middleman.
Whereas on IPFS files are stored an the nodes, these nodes keep only the files that are relevant to them ie.how many times have they been requested by the nearest peer. The major difference on security is that IPFS files are requested by their cryptographic hash so none has any idea as to what might these files contain. Not the nodes nor the middleman.
Only the person making the valid request can view the content.
This is the reason IPFS websites or files are not accessible by Chrome etc. without the addition of a plugin or extension.
I hope this solves your query about ISP requirement for IPFS and Encryption over IPFS.
The part where IPFS can’t be taken down is true in the sense that there has to be at least one node which hosts your data, the same way anytype is hosting a caffe node for each alpha user currently to keep their data synced. Right now IPFS is a rapidly emerging technology the underlying idea is that one day there will be so many devices on the IPFS network that it would be virtually impossible to render all these devices useless and disable IPFS, also these devices will be able to host everyone’s data in multiple nodes for redundancy.
Hope this helps.