Notes from "Structuring IPFS Research" discussion

[jbenet]

Some of us had a discussion about structuring IPFS research, taking learnings from how Filecoin has done its research, and PL's RFPs.

Filecoin Research links:

• https://github.com/filecoin-project/research
• http://research.filecoin.io

PL Research RFPs

• https://github.com/protocol/research
• https://github.com/protocol/research-RFPs

---

Research Topics

These are some of the areas to prioritize. By no means comprehensive or complete. These are quick notes.

Security

• Security Breaks
  • how do we handle breaks in primitives
• how do we handle cryptographic artifact rot

IPLD

• mapping hashes large blocks to smaller blocks

• can we use snarks to map large blocks to smaller blocks

• big data

• verifiable compilers

• reproducible computation

Naming Related

• Key names/mutable pointers
• security threat model
• Networking transports
  • How to be flexible?
  • DHT vs Pubsub
• Does naming affect routing?
• Transport and security properties varying vs human expectations

Mutable Data

• addressing single mutable references
• addressing groups of references
• multi-writer mutable pointers

Networking Related

• Content Routing

  • DHTs
  • lower latency
  • Privacy Preserving
  • Oblivious Content Routing
  • Oblivious Traffic Routing
  • Scalable and efficient random access into large datasets
  • Trillions of object
  • Hierarchical / multi level DHTs
  • topic based DHTs
  • Content Routing SLA

• Multicast / Pubsub

  • Efficient propagation of things

• Machine learning for networking

  • topologies
  • congetion control
  • TCP Remy style things
  • tuning parameters
  • censorship evasion

[nicola]

IPLD

• Alternative Vector Committments/Accumulators to Merkle tree
• Better IPLD selectors based on better Authenticated Data Structures

[miyazono]

I moved all this to a google doc for async discussion here. I'll link conclusions when I move them back to github.

[tapaswenipathak]

Hi folks: Can I read the doc? or is the discussion internal? Does the doc has more than the ticket?

[miyazono]

hey @tapaswenipathak, thanks for your interest and for following up. I think the doc has some incomplete internal discussion at the moment, but we'll definitely be moving some of the content back to github.

Tagging some people here for awareness: @daviddias, @jsoares

[tapaswenipathak]

Hey @miyazono, np, thanks for the info. :)