New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
IPIP-252: efficient content-routing proposal for huge IPLD trees #252
base: main
Are you sure you want to change the base?
Conversation
@@ -0,0 +1,264 @@ | |||
# ![](https://img.shields.io/badge/status-wip-orange.svg?style=flat-square) Content routing | |||
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
suggestion - start with a short recap or set of references to the status quo to orient new readers (landing on problems raises the questions "problems with what" if no background is given initially)
|
||
Remark: The IPFS codebase has an interface named `ContentRouting`. However, this is an interface to low-level DHT functions rather than to a wholistic content routing system as we describe here. | ||
|
||
For instance, when bitswap fetches the contents of a CID, first it tries to find a provider using gossip-based queries to peering bitswap nodes and, if this fails, it tries to find a provider record in the DHT. This can be viewed as using two independent content-routing systems. Let's call them DHT-based and gossip-based, respectively. The DHT-based system has high latency and high success rate. Whereas the gossip-based system has lower-latency and lower success rate. To get the best of both, the DHT-based system is used as a fallback, should the gossip-based one fail. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
the "respectively" is vice versa, right? peered bitswap nodes is gossip-based.
|
||
## Lack of abstraction | ||
|
||
"Content routing" loosely refers to the mechanism that enables some nodes to advertise the data they have, and for other nodes to discover it, by a content identifier (CID). Currently content routing is not a unified software subsystem of IPFS, instead it is a workflow (involving using the DHT and other methods) that various applications (e.g. bitswap and graphsync) in need of content routing replicate individually and vary according to their need. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Suggest rewording last sentence to:
Currently content routing is not a unified software subsystem of IPFS, but is instead a workflow that uses the DHT and other methods. This workflow is replicated by various applications (e.g. bitswap and graphsync) and is varied according to the application's content routing needs.
No description provided.