A few short talks introducing the IPFS þing week, how it will work, & what to expect.


IPFS þing Opening

Meet & Greet

Welcome & Introductions

IPFS þing Orientation and Q&A

Kickoff Keynote

There is a hackspace set up in the common area.
Tables, couches, power, drinks, and wifi.


Hack, Cowork, Discuss

There is a space to schedule sessions that emerge out of other conversations or for sessions that don't fit other places.
More details for scheduling are in the IPFS Discourse topic: https://discuss.ipfs.io/t/unconf-and-miscellaneous-sessions/14676


Unconf and Miscellaneous Sessions

This track provides an overview of IPFS implementations. Each presentation is about a different IPFS Implementation, and meant to be a pitch for the implementation to prospective users, contributors, etc. Think of these presentations as listed in a webpage about IPFS implementations.

Each talk should cover:
- What the implementation is for (problems / solutions)
- How it works / architecture
- Team (who's building it)
- Traction so far
- Roadmap
- Painpoints: what the impl is not good at.

(?) - denotes tenative (not confirmed)


IPFS Implementations

Lessons from putting IPFS in weird places

Kubo (go-ipfs)

Gateways in the past, present, and the future

IPFS Gateways

IPFS Companion, IPFS Desktop, and Brave

The WebNative File System adds versioning, cryptree-based encryption, and eventual consistency on top of IPFS

WebNative File System (WNFS)

Introducing Iroh: Performant IPFS for the whole world, right now

Iroh

IPFS Cluster is an P2P application that uses the kubo (go-ipfs) menu with some spices on top

IPFS Cluster

Break

Filecoin -- an exabyte scale IPFS system

Lotus - an IPFS blockchain node

An brief overview of the read and write pipelines in 
nft.storage and web3.storage implementing IPFS protocols across several 
services and infra providers

Web Scale IPFS

Estuary

IPFS Operator

Auspinner is a CLI tool for pinning CARs to Pinning API Spec compliant services using Bitswap.

Auspinner

libipfs is an upcoming library of IPFS components for building IPFS implementations

libipfs

js-ipfs

Aqua is a π-calculus based language for scripting decentralized protocols on top of libp2p, IPFS and beyond.

Aqua language

Peergos builds on IPFS to add quantum-resistant, metadata-protected encryption with Cryptree+, block-level access control with authed-bitswap, and a super minimal ipld daemon with ipfs-nucleus, whilst maintaining autoscaling and efficient modification properties of IPFS.

Peergos private filesystem

If we can establish subsecond on average resolution across any CID, that would open a massive amount of adoption across IPFS


Content Routing 1: Performance

Welcome standup discussion

Content Routing Intro: Performance

@guseggert on IPFS DHT, @gammazero on Network Indexer, @ajnavarro on Reframe

State of content routing lightning talks

Open Problem Identification and Alignment

Domain-specific IPFS content indexer with Aqua

Lunch Break

Content Routing - Scalability Constraints

Hierarchical Consensus for Content Routing

Network indexer roadmap

How to be an index provider

An even deeper dive specifically into how Elastic IPFS interfaces with the Indexer Nodes. This focuses on the specific advertisment mechanism Elastic IPFS uses to get data into the Indexer Nodes (i.e. HTTP+written advertisments in S3 bucket). There are a few, so it is interesting to see a specific method in action. We'll also lightly cover how the indexer nodes are part of the delegated routing protocol, how the Hydra nodes are currently the consumers of this interface and why that matters.

How Elastic IPFS content is discoverable in the DHT

Paths to Decentralize Network Indexers

Offering strong privacy properties in content routing is a critical concern for decentralized networking.
This track will cover current efforts to improve content routing privacy, and potential opportunities to do even


Content Routing 2: Privacy

Content Routing Track Intro

Secure & Private Content Routing

Private Set Intersection for Bitswap

Building towards a secure and performant system

Funding Private Retrieval

Mixnets & Content Routing

Double Hashing & Content Routing

Use cases for privacy in LibP2P

Private Information Retrieval & Content Routing

Open Problem & Project Generation Workshop


Part 1 of the "Data and IPFS" track will cover pain points stemming from the way data is structured within IPFS & IPLD

- Making UnixFS not special (e.g. rendering other types of files/folders on gateways)
- The relationships between Checksums, Blocks, etc.
- Compatibility with existing (and new) hash linked data structures
- Abstractions we have and want for working with hash linked data structures

People should come prepared with a prompt for "What's your vision for working with data in IPFS, and what about how it works today causes you the most pain?".
Write it up in https://pl-strflt.notion.site/Your-vision-for-IPFS-and-Data-15087c9885db48df8af4f11fc62d15e1. See the welcome discussion for slot for more info.



Data and IPFS: Models

A track introduction and discussion to figure out:
1. Why each of us are here and the backgrounds/use cases motivating us
2. Introduce an overview of our shared problem space
3. Identify what we want to accomplish during the day

People should come prepared with a prompt for "What's your vision for working with data in IPFS, and what about how it works today causes you the most pain?".
Write it up in https://pl-strflt.notion.site/Your-vision-for-IPFS-and-Data-15087c9885db48df8af4f11fc62d15e1.


Welcome standup discussion and track intro

The 2MiB block limit is a major source of interoperability pain, why does it exist and how can we deal with it?

Dealing with block limits

This talk will go over how IPLD could be exposed as a first class citizen in gateways and how we can represent links to data with URLs.

IPLD on gateways, and with URLs

Within IPLD we have a variety of abstractions like codecs, schemas and ADLs. What do we have and what do we need?

Group Discussion: Abstractions for hash-linked data

The WebAssembly Codec was designed to serialize and transport IPLD Data Model data for the purposes of building IPLD tooling in WASM.
This talk will describe today's IPLD Data Model as well as how WAC is used to represent it.


What's with the IPLD Data Model, and what happens when you try and represent it - WAC

UnixFS has been an IPFS staple for a long time, but it doesn't solve all the use cases people have for describing filesystems. What changes do we want to make and how would we make them?

Group Discussion: UnixFS and what comes next?

We have an afternoon of open data modeling and transfer slots tomorrow, what areas are ripe for new or continuing discussion?

Recap and brainstorming

Part 2 of the "Data And IPFS" track will focus on content-addressed data transfer protocols, focusing on four key areas:

1. Multi-party data transfer
2. Efficient graph synchronization
3. connection integrity
4. connection saturation



Data And IPFS: Transfer

This is track introduction and facilitated discussion to:
1. Find out why we're here
2. Help folks get shared understanding on the overall problem space
3. Identify what we want to accomplish during the day


Welcome to the world of data transfer

stebalien is one of the keepers of 'human spec' on how bitswap works in practice. In this talk he'll give a brain dump of bitswap, and do a Q & A on the details

bitswap AMA

Hannah will give an overview of Graphsync—how it works, what it can do in 2022. 
We'll also cover some challenges and why it is hard to make GraphSync work well.


Why everyone hates GraphSync, but maybe shouldn't

CAR Mirror, CAR Pool, and Private-XOR are attempts to radically reduce the number of round trips, minimize (but not eliminate) redundant blocks, and push and pull data over mature channels, in unicast, multicast, and anycast scenarios. This is well suited to undialable nodes such as browsers, directly connected swarms (e.g. after DHT lookup), and broadcasting blind indices.

CAR Pool and more

This talk will define a common pattern found in graphsync, bittorrent, dsync, and CAR files: clients fetching an index of the content to optimize data transfer. We'll examine the benefits & drawbacks of this approach, and connect this back to the architecture of IPFS

What's in a manifest?

Let's talk about what we've learned and next steps

Group Discussion: Data Transfer


This track aims to create space for advancing the IPFS project and community. New session proposals are welcome, especially those that focus on a specific, actionable subtopic rather than attempting to grapple with the entire universe of project and community stewardship at once.



Project & Community

A walkthrough of the new IPIP (RFC-like) process and its motivations, updated specs repository, and when and why you should submit an IPIP. Q&A at the end.

InterPlanetary Improvement Process (IPIP)

Discuss evolving web-wide standards & specs that might benefit from participation from our community. Examples include upcoming deadline at IETF for multi* stuff started long ago, and W3C activity around extensions APIs and webviews and wallet APIs.

Shaping IETF and W3C Standards

IPFS has 'won'. Content addressing is a sustaining innovation that can make its way into many more places. Discuss creating a cross-organization Content Addressable Alliance that advocates and promotes the usage of content addressing.

Content Addressing Alliance

Brief session on history of funding IPFS implementations, and exciting updates going forward.

Funding IPFS Implementations

Learn what happens in a given month in IPFS development for projects like Kubo, js-ipfs, and supporting the wider community.  You should walk away with better insight into venues you can engage the maintainers and ideas/input will be welcome as well.

A month in PL EngRes IPFS Development and how you can be involved

This session will be a walk through how PL manages IPFS GitHub organization configuration as code using a tool called GitHub Management. It'll provide practical information on how to request changes to the configuration (for example, if you contribute frequently to a repository within IPFS and you'd want to become a collaborator) and how to set up a similar solution in any other GitHub organization.

GitHub Configuration (of IPFS organization) as Code

Mutual Therapy Session for operators, maintainers, founders, or ceos

**New directions for scalable virtual worlds.**

A short workshop for IPVM planning.



IPVM Planning

Planning workshop for IPVM

Connecting IPFS protocols with other decentralized protocols to 
meet user needs, often by building and operating infrastructure that brings them together.
Talks in the morning, then an hour long group discussion on 'Writable Gateway' followed by Unconf.

Connecting IPFS

Track introduction

Walk through and demo of how the w3name service works currently

w3name: IPNS republishing as a service

Iroh Cloud overview

Policy based content providing for IPFS peers

IPFS safemode tool

Caching IPFS & Bitswap Workers

A run through of Elastic IPFS as it has been implemented in AWS: an in-depth discussion of the various sub-systems of Elastic IPFS, explaining the function(s) of each, the choices made and why, and the trade offs.

The Elastic IPFS implementation in AWS

The journey from initial working implementation to production system. A short story of metrics and data in graphs.

The Rise of Elastic IPFS

Connecting IPFS users and permissions to DIDs with UCAN

This talk will look at the existing state of writable gateways and some potential paths for making them useful for browsers and applications in general

Writeable Gateways Discussion

Unconf


Part 3 of the "Data and IPFS" track is an unconference for any topics related to
either data models or data transfer on IPFS.

We may also schedule follow-ups to topics covered in either Part 1 or Part 2


Data and IPFS: Unconf

Review what we has already been discussed in parts 1 and 2, and pitch topics.
Pitch a 10-15 minute discussion or claim two slots for 30 minutes.


Intro and Afternoon Planning

Open Slot

Group discussions summarize learnings for everyone. Review overall learnings from
parts 1, 2, and 3 and discuss next steps


Report backs and closing

**The goal of this track is to craft a vision of IPFS in the web and browsers which is truly
 native to *both*, and a roadmap for achieving it.**

The web platform itself is a massive heterogenious ecosystem with unique challenges for new protocols (something which has never happened in a meaningful way) and stakeholders with particular requirements and a diverse vision of what the web actually is and can be. We need to define a way for IPFS be a deployment channel for HTML/JS/CSS, have a clearly defined security model for that content, such that both web standards and browser vendor stakeholders can reason about it and ultimately implement it.

Additionally, extra-web browser features and extensibility channels like browser extensions offer different capabilities, audiences and constraints. We must determine what the optimal strategy is for augmenting user experience of the regular web through the opportunities we have which are *adjacent* to it - eg connecting to local IPFS nodes, building browser extensions which read from or write to IPFS, or enabling users to move HTTP content to IPFS.

Finally, looking beyond existing browser channels, how can we redefine or leapfrog today's web using entirely new approaches such as removing existing browser metaphors or building mobile operating systems from the ground up with IPFS at the center.

This track is structured as an _intense_ half-day workshop. It will start with a
 brief introduction to frame our work, we'll hear lightning talks on use cases, 
then move to group discussion, using a design document for group iteration.

Bring your ideas, use-cases, needs and we'll redefine the track as needed before kicking off.


Browsers and The Web Platform

What's happened so far, who are the players, what's the landscape, 
what barriers exist, what feature gaps need closing, what to do next.

Intro to IPFS in Browsers & the Web

State of IPFS and HTTP (TBD)

This talk will look at Peer to Peer URLs and how they can be used
in tandem with HTTP-like APIs within webapps using custom protocol handlers.

Agregore `fetch()`, a multi-protocol abstraction

Incumbent Mobile OSes centralize content discovery in their app
stores. Let's explore how we could change that with IPFS!

IPFS and Content Discovery

Manifest V3 and the road forward for IPFS browser extensions

Will provide a quick overview of Webrecorder tools and how they support saving web sites using the browser, from a single page to many TBs of data, and accessing high fidelity web archives from IPFS. Will cover ongoing work, research and challenges.

Saving the web to IPFS while you browse with Webrecorder tools

While full nodes can use transports like TCP and QUIC, this is not
an option for web browsers. Instead, we have to fall back to protocols like WebSocket,
WebTransport and WebRTC. And even for those, the browser API severly limits what we
can do.

Challenges in Browser Connectivity


# IPFS and WASM track

Come and discuss how to imbue IPFS with WASM magical powers.

(?) - denotes tenative (not confirmed)



IPFS & WASM

Track Intro

We have wasm-ipld code that can run in kubo and work with gateways. Let's chat about it and what comes next

wasm-ipld demo and discussion

FVM Deep Dive

WASM + IPLD

WASM forces us to build pure interfaces that keep the door open for different networking strategies.

Using IPFS/IPLD from WASM

Wasm modules, their arguments, intermediate states, their outputs, and managed effects can be described as IPLD graphs. IPVM is a strategy to support generalized deterministic computation in a serverless style on top of IPFS with optional sidechannel matchmaking on Filecoin, and extend the same benefits of shared data blocks to computation.

IPVM

Autocodec links IPLD graphs to Wasm modules that can decode them. This has applications for dynamically loading codecs, optimizing popular codecs, and sharing decoded blocks directly

Autocodec

Fast demo demonstrating rust-libp2p running inside the browser with Webassembly

Rust-Libp2p-WASM Demo

How can a WASM-powered IPFS/libp2p node help with interoperability and deployment of new protocols

Idea/Brainstorm: Implement once, run everywhere

VMs & IPFS - use cases

This talk will dive into details of the Wasm Component model and how we approach it in Fluence Labs for services

Marine: dynamic Wasm modules linking

Planning Discussion

A discussion / breakout-oriented workshop for defining and committing to next steps out of the week's conversations, which we can land and celebrate at upcoming IPFS events in Q3 / Q4 2022.


Roadmapping Next Steps out of the IPFS þing

A sequence of closing keynotes summarizing the week and
looking ahead to the rest of 2022 and 2023.


IPFS þing Closing


From Developer Experience to User Experience, how do we build full apps on IPFS? What are the goals of IPFS beyond infrastructure?

Select events to join and follow in the IPFS Discourse https://discuss.ipfs.io/t/2022-07-13-building-apps-on-ipfs-track/14605



Building Apps on IPFS

Can we make IPFS the protocol a default deployment target like Vercel or Netlify? What does apps on IPFS even mean? Does data portability itself on IPFS something that can be added to any apps? Open discussion, bring your thoughts! Track intro event is on the main notes page https://discuss.ipfs.io/t/building-apps-on-ipfs-track/14605

Track Intro: Building Apps on IPFS

9:30am - 11am. @expede - Fission Webnative SDK: DIDs, UCANs & WNFS, @ianopolous PeerGos architecture, @smrz2001 Ceramic, CACAO, DID Session Keys. 3 short 15min overviews followed by 45min of discussion on working together.

Identity, Capabalities, & Private Data

Subconscious is a tool for thought with the goal of fostering a credibly unowned, globally-distributed knowledge graph. Noosphere is the IPFS-adjacent protocol that drives Subconscious. In this session we will discuss the opportunities and challenges that led us to using IPFS, and detail some of the designs we are implementing to serve requirements in naming, authz and network availability. Event on IPFS Discourse https://discuss.ipfs.io/t/subconscious-noosphere-www-over-ipfs/14652

Subconscious & Noosphere: WWW over IPFS

this talk will explore tensions in the design of IPLD, multicodecs, and UnixFS, and an opportunity to open up a new class of application development on IPFS by re-orienting toward arbitrary metadata & a WASM runtime

Application space in IPFS

The Compute Over Data Working Group is just getting started. We'll have intros from WG members on their compute solutions and how they intersect with IPFS, and how they see powering appss.

Compute + IPFS

Come together and review all issues and points of discussion harvested during the day. Assign follow ups, working group creation, and other connections as needed (aim for 30min)

Wrap Up Session


Aqua is a programming language and a runtime for distributed algorithms based on π-calculus. Aqua creates secure, ad-hoc overlay networks for every execution making easy to iterate on deploying, testing, and upgrading permissionless p2p protocols.

Come to learn about Aqua and discuss with us how it can help improving IPFS.



Aqua and IPFS

Aqua language intro

Workshop. Domain-specific IPFS content indexer with Aqua

AquaVM: calculus based VM for distributed algorithms

How Aqua can simplify and enhance IPFS/libp2p protocol development

Idea/Brainstorm: Aqua and IPFS

_You can’t improve what you don’t measure and you should measure what (you think) you’ve just improved._

We need to get to a point where we can measure the performance of all protocol and system components of the IPFS stack, identify bottlenecks and propose optimizations.

In this track we will present the measurement studies we've kickstarted and will then go on to explore the components that will benefit from more detailed monitoring. We'll also workshop on the details of the methodologies needed in order to actually have this monitoring in place. Ultimately, the session's target is to identify how we can transition towards a culture of protocol measurement, monitoring & optimization.

Anyone is welcome to pop in and bring new items to our attention.


Schedule