mirror of
https://github.com/foo-dogsquared/wiki.git
synced 2025-01-31 10:58:28 +00:00
b088086b06
Now, it's all under the notebook umbrella. Seems to be appropriate as it is just my notes after all. I also updated some notes from there. I didn't keep track of what it is this time. Something about more learning notes extracted from my "Learning how to learn" course notes and then some. Lack of time and hurriness just makes it difficult to track but it should be under version control already.
24 lines
1.5 KiB
Org Mode
24 lines
1.5 KiB
Org Mode
#+title: IPFS protocol
|
|
#+date: "2021-05-13 16:41:26 +08:00"
|
|
#+date_modified: "2021-05-13 16:41:36 +08:00"
|
|
#+language: en
|
|
|
|
|
|
A project that aims to power the next generation of the web (Web 3.0).
|
|
The core component is a protocol (also called IPFS protocol) that aims to replace HTTP.
|
|
|
|
In order to understand what sets IPFS apart, we need to discuss a bit on the current way of delivering content.
|
|
|
|
Currently, 99% of the web today are delivered through HTTP.
|
|
It is primarily delivered through *location-based addressing*.
|
|
That is, in order to get what the client needs, they need to send a request to the server that has that content wherever they are.
|
|
[fn:: Yes, content delivery networks (CDN) exist but we'll disregard it for now since it is not baked in on the protocol.]
|
|
Problems can come through if the server is located in another country or continent, even, which will result in slower requests.
|
|
Furthermore, it will create a bottleneck for that server especially if it's a popular site like Facebook, YouTube, and Twitter.
|
|
|
|
IPFS addresses this problem by settling into a different approach: *content-based addressing*.
|
|
In this protocol, when the client requests for the content, it will crawl through the network searching for the node that has the content.
|
|
|
|
For content-based addressing, they make content hashes as the central component for that as a way to identify the content.
|
|
But since cryptography is not perfect and can always be broken, content IDs can come in multiple format.
|