mirror of
https://github.com/logos-storage/logos-storage-research.git
synced 2026-01-25 00:33:10 +00:00
150 lines
2.9 KiB
Markdown
150 lines
2.9 KiB
Markdown
---
|
|
title: Dagger
|
|
theme: solarized
|
|
revealOptions:
|
|
controls: false
|
|
progress: false
|
|
transition: 'fade'
|
|
---
|
|
|
|
# Dagger
|
|
|
|
Decentralized Storage
|
|
|
|
Note:
|
|
Briefly describe what Dagger is.
|
|
|
|
---
|
|
|
|
|
|
## Why Dagger?
|
|
|
|
 <!-- .element: width="30%"-->
|
|
|
|
Note:
|
|
Decentralized storage is a fundamentally unsolved problem.
|
|
P2P storage hasn't taken off, why?
|
|
Without reliable storage, dapps won't take off.
|
|
Needs same guarantees as cloud storage, and then some.
|
|
|
|
---
|
|
|
|
## Reliability
|
|
|
|
predictable
|
|
|
|
accessible
|
|
|
|
Note:
|
|
Reliability is the property that is most lacking in p2p storage networks.
|
|
Predictable storage duration.
|
|
Storage is always accessible to you.
|
|
|
|
---
|
|
|
|
## Proofs
|
|
|
|
data possession
|
|
|
|
retrievability
|
|
|
|
|
|
Note:
|
|
How do we achieve reliability?
|
|
Proof of data possession over time
|
|
Proof of retrievability (interactive)
|
|
|
|
---
|
|
|
|
proof of data possession
|
|
|
|
 <!-- .element: width="60%"-->
|
|
|
|
Note:
|
|
Proof of storage.
|
|
File is divided into chunks.
|
|
Over time, proofs of different randomly selected chunks are requested
|
|
|
|
---
|
|
|
|
proof of data possession
|
|
|
|
 <!-- .element: width="60%"-->
|
|
|
|
Note:
|
|
We improve on this scheme by making sampling events dependent on each other.
|
|
This increases the odds of detecting missing chunks over time.
|
|
|
|
---
|
|
|
|
proof of retrievability
|
|
|
|
 <!-- .element: width="70%"-->
|
|
|
|
Note:
|
|
Prevents data withholding (ransomware)
|
|
The prover does not know when it's being verified
|
|
|
|
---
|
|
|
|
## Incentives
|
|
|
|
storage
|
|
|
|
bandwidth
|
|
|
|
 <!-- .element: width="60%"-->
|
|
|
|
Note:
|
|
Incentives compensate for resources
|
|
Incentives provide security; it makes economic sense to stick to the rules
|
|
Storage incentives provide storage reliability
|
|
Bandwidth incentives provide retrievability (dynamic CDN)
|
|
Incentives encourage participation in the network
|
|
|
|
---
|
|
|
|
## Money / Investments
|
|
|
|
Note:
|
|
how much of the market can we cover
|
|
how much would it cost to make
|
|
|
|
---
|
|
|
|
## Privacy & Anonymity
|
|
|
|
 <!-- .element: width="40%"-->
|
|
|
|
Note:
|
|
Dagger layer: identities used for incentives and payments are pseudonymous.
|
|
Libp2p layer: peer identity is pseudonymous.
|
|
IP layer: depending on the need, there are several solutions to choose from
|
|
|
|
---
|
|
|
|
## Network Overview
|
|
|
|
 <!-- .element: width="65%"-->
|
|
|
|
---
|
|
|
|
## Software Stack
|
|
|
|
 <!-- .element: width="70%"-->
|
|
|
|
---
|
|
|
|
## Research agenda
|
|
|
|
- proofs for large data sets
|
|
- formal/emperical evidence
|
|
- consensus engine
|
|
|
|
Note:
|
|
Small proofs, constant size, aggregatable, computationally light
|
|
Zero knowledge proofs
|
|
Formal proofs/emperical evidence that our algorithms have the desired properties
|
|
Model the network to observe emergent properties
|
|
How does the storage network interact with a consensus engine
|