- This site is published to https://ipfs-shipyard.org/ using https://super.so/.
- Credentials are in 1password.
- Publishing happens automatically.
- Any content in a brown toggle like this is hidden with CSS in publishing (but is still in the DOM). This has been our quick hack for how to have content that doesn’t show up on the published website.
- The current “IPFS Shipyard” logo comes from https://www.bing.com/images/create/create-a-logo-by-taking-aan-ipfs-hexagon-and-putti/1-654f00e96ced4a509c17899cd5a5b3d3?id=g5GvBZkYkbJ7xdClAC5wRw%3D%3D&view=detailv2&idpp=genimg&FORM=GCRIDP&ajaxhist=0&ajaxserp=0
Since 2014, IPFS has been used to distribute exabytes of data across the globe, with 250k+ public p2p nodes, and over 11M weekly users to HTTP gateways.
We’re thrilled to see many projects in both Web3 and Web2.5 powered by IPFS. These include dapp frontends like Uniswap, Audius digital music streaming, dClimate’s decentralized marketplace for climate data, Anytime’s rich desktop application for content creation, and Gala’s game asset distribution. IPFS development was founded and forged in Protocol Labs, but the project is maturing with:
- Establishing an independent IPFS Protocol Foundation
- Graduating a core engineering team with over 25 years IPFS development experience from Protocol Labs.
We, the IPFS Shipyard team, will be an independent engineering team running hard at making the next 1 billion websites and web apps faster, resilient, and more open.
How do we make IPFS & IPFS-based tools the first choice for anyone publishing apps or data on the web?
- Remove any overhead of retrieving and serving content addressable bytes by leveraging existing web infrastructure for moving and caching bytes with HTTP.
- Add world-class support for the existing content addressable data with support for large blocks.
- Make Kubo and Boxo SDK radically more developer-friendly, with features chunked into better abstractions.
- Develop better tools for operating IPFS with request tracing, debugging, and denylist handling.
To underpin this all, we’ll continue to maintain key common goods software and services for IPFS such as Kubo, Helia, the Amino public DHT network, and the ipfs.io (+dweb.link/trustless-gateway.link) HTTP gateway.
In 2024, we’ll focus on web publishing for decentralized apps. This has been a key use case since the beginning of IPFS, but the current stack includes many compromises and inefficiencies. Today, a small slice of the most avid IPFS users download and install a browser or extension that can speak the IPFS protocol directly. For everyone else, requests get routed through HTTP gateways. They still need HTTP and centralized servers, with all the associated problems.
We can do better. Our team is making tremendous progress on a 4-step IPFS web fetching progression. Trustless gateways, delegated routing, and service workers (in a coordinated series of changes across Kubo, Rainbow, Helia, and public network infrastructure) will push more p2p work to the browser, at no cost (i.e. no installs required) to the end user.
Our seasoned team has over 35+ combined years in the IPFS ecosystem as maintainers, specification stewards, and technology builders. We have the technical depth, experience, and management to work across the stack making protocols and supporting tools in the browser and backend systems.
By the end of 2024, gateways will be radically cheaper to run and webapps will meet the ESMA definition of “fully decentralized”. [The European Securities and Markets Authority (ESMA) recently published a paper stating that fully decentralized frontends are exempt from their scope of regulation. Meeting that standard will also make the public network more resilient to outages and censorship.]
Building this new reality, and continuing to maintain core common goods and services, will cost ~$4M over 12 months. We think this brings tremendous value to many projects and companies across Web2.5 and Web3, including [yours].
Protocol Labs will be a long-term, major sponsor of this work, but it’s time to bring in the broader IPFS community. We can’t do this without you. We’re seeking grants and support from a small number of mission-aligned partners.
❓ If your business or endeavor…
…depends on projects like Kubo, Boxo, or Helia being well-maintained and developed OR
…relies on a healthy public DHT OR
…benefits from accurate and evolving IPFS specifications
Please add your financial support (or contact us to discuss and learn more)
100% of your grant (minus fund administrative fees) will go to the maintenance, development, and operation of our projects by enabling our exceptional team to stay dedicated to these efforts. (Crypto and FIAT welcome.)
❓ Is your company…
…building on IPFS, and you’ve been hacking around issues in Kubo, Helia, etc.?
…interested in using IPFS, but after your initial demo, you’re not sure where to get started?
Reach out for a free consultation exploration for us to bring our expertise to help navigate your journey.
This could include fixing problems in underlying implementations or libraries, or helping build tooling that enables solving the problems with you. Engagements here can be one-time in nature for things like architecture consulting or bug fixing or feature development. They can also be recurring with supporting certain features long term or helping shepherd ideas through specs/governance processes.
Our consulting arm and non-profit entity structure to collect donations will be established by the end of 2023, but your verbal commitment in the interim is very helpful. Please read more or
Answering how much data has been transferred with IPFS is actually a hard question to answer. Two easy large data transfer pots that can point to that are measured are:
- Filecoin with its 1.6 EiB of data committed on chain in 202310 per https://dashboard.starboard.ventures/market-deals
- ipfs.io gateway which transferred 2 PB of day in the last 30 days per https://protocollabs.grafana.net/d/J2_IHYTVz/gateway-report?orgId=1&viewPanel=20&from=now-30d&to=now (which isn’t public but we’ll get it so in 2024!)
Getting unique client IP addresses in a given week requires looking at some internal dashboards.
https://protocollabs.grafana.net/d/J2_IHYTVz/gateway-report?orgId=1&editPanel=54&from=now-30d&to=now
The public daily unique IP addresses shows similar results: https://www.probelab.io/ipfsgateways/#gateway-clients-overall-plot
Artwork originally comes from https://www.explainxkcd.com/wiki/index.php/2347:_Dependency
It was then modified in https://docs.google.com/drawings/d/14gif1V200f0_WMlV11IWB39jW8LksufYvca8dJy7izs/edit
Since 2014, IPFS has been used to distribute exabytes of data across the globe, with 250k+ public p2p nodes, and over 11M weekly users to HTTP gateways. IPFS development was founded and forged in PL Inc, but the project is maturing with:
- establishing an independent IPFS Protocol Foundation (public blog post coming)
- graduating a core engineering team with over 20 years IPFS development experience from Protocol Labs Inc.
We, the
How do we make IPFS & IPFS-based tools the first choice for anyone publishing apps or data on the web? Large-scale data handling needs to become world-class, starting with support for large blocks and HTTP retrieval. Kubo and Boxo SDK needs to be radically more developer-friendly, with features chunked into better abstractions. Operating IPFS needs better tools in the areas of request tracing, debugging, and denylist handling.
To underpin this all, we’ll continue to maintain key common goods software and services for IPFS: Kubo, Helia, the Amino public DHT network, and the ipfs.io HTTP gateway. We’re thrilled by the content addressing projects that have sprung up around us over the years, and we feel the responsibility to keep our areas of stewardship solid while also empowering developers to more easily help themselves in new ways.
PL Inc will be a long-term, major sponsor of this work, but it’s time to bring in the broader IPFS community. We’re excited for more transparency and community participation, and that’s where you come in. While we have historically aligned our team’s priorities with PL Inc.’s wider initiatives, this new phase allows us to be more responsive to the needs of our direct users and even contract our services and expertise. It’s time for community having greater input on development, and that comes with supporting it financially.
❓ If your business or endeavor…
…depends on projects like Kubo, Boxo, or Helia being well maintained and developed OR
…relies on a healthy public DHT like Amino OR
…benefits from accurate and evolving specifications
would you consider sponsoring our development?
100% of your sponsorship (minus fund administrative fees) will go to the maintenance, development, and operating of our projects by enabling our exceptional team to stay dedicated to these efforts. If so, we’d love to hear from you, including any input on functionality you’d like to see prioritized on our roadmap.
❓ Alternatively, or in addition, are you a company…
…building on IPFS and you’ve been hacking around issues in Kubo, Helia, etc.?
…interested in using IPFS, but after your initial demo you’re not sure where to get started?
If so, would you consider hiring us to bring our expertise to bear to help navigate your journey?
This could include fixing problems in underlying implementations or libraries or helping build tooling that enables solving the problems with you. Engagements here can be one-time in nature for things like architecture consulting or bug fixing or feature development. They can also be recurring with supporting certain features long term or helping shepherd ideas through specs/governance processes.
Our consulting arm and non-profit entity structure to collect donations will be established by end of 2023, but your verbal commit in the interim is very helpful. Please
- Reliable retrieval
- When helia is given a CID, it can get the content whenever it’s located
- Reach out to more powerful nodes like gateways in a trustless fashion via trustless gateway API, with verification on your side.
- More p2p work within the browser
- Find the right other peers to connect with
- Basically free for anyone running a gateway to turn on delegated routing, decreasing the load on the gateway operators. delegated-ipfs-tech. also reduces likelihood of DMCA blockage. you now have a means to more efficiently find peers. across firewalls - libp2p webRTC connectivity across firewalls (coming in next release)
- Gateways as true transitions
- Today, gateways are a cheat code. They make IPFS or even, but they’re very expensive to run.
- The new Helia Service Worker Gateway
- When you add,
- More resilience
- Add delegated routing
- Gateways
- This means a radical reduction of infrastructure costs for anyone running a gateway.
- Benefits:
- More true decentralization
- More reliable/resilient — match or exceed HTTP
- Gateways will get cheaper to run
- Example: ipfs.io is $1M/year to run
- New: each IP address gets 500M/week, beyond that you need to install a service worker, and the rest of your bandwidth has to go p2p.
- If dapp frontends are paying now to host their sites, it would reduce opex.
Sketches of more notes to weave in
Option 1:
Like rust out Rust out of Mozilla
Support upcoming Foundation
Option 2:
Link to a protocol.ai blog post.
Good words
- Getting independence
- Graduated
- Participant in the “innovation network”
Other ideas to include:
- Lean into easier to influence