If there’s one thing that the web3 community seems to be in agreement on, it’s a love for IPFS. NFTs, web3 social, DAO’s, protocols, and infrastructure projects alike are all leveraging the distributed file system for a multitude of use cases.
We should know, Fleek is one of those projects ⚡ After years of building in the IPFS ecosystem, hosting tens of thousands of web3’s apps on IPFS, and being one of the largest pinning providers, we’ve learned a lot!
Today, we’re outlining the multiple ways that Fleek Network, the decentralized edge & content delivery network, will help to decentralize IPFS, especially as it relates to IPFS pinning and IPFS gateways.
The Current Problem With IPFS Pinning & Gateways
Currently, there is no easy and trustless way to ensure that your content on IPFS is always available. With centralized pinning providers and gateways, you’re forced to run the risk of your content falling off IPFS and becoming unavailable, suffering gateway downtime, or being forced to store files twice because web3 storage protocols don’t integrate with IPFS directly.
Because of this, IPFS right now is pretty centralized and prone to censorship. The majority of the IPFS network and usage is facilitated and powered by a small handful of known corporate entities (ex. ‘pinning providers’). That’s not very web3, and given the universal love and widespread usage of IPFS, we believe it’s an important problem to solve. Fleek Network will significantly alleviate these issues.
How Fleek Network Helps Decentralize IPFS
Giving all Storage Layers the Benefits of IPFS
Fleek Network will help to further decentralize IPFS by acting as a connective layer that sits on top of all storage protocols and trustlessly connecting the benefits (content routing and addressing) of IPFS to them. This will make all files addressable by a single CID, instead of treating them as a separate and redundant cold storage storage layer only accessible through custom logic/routing.
That is because Fleek Network nodes are also IPFS gateways, and can fetch data from IPFS even if the files aren’t cached on Fleek Network yet. The only difference between a Fleek Network node and an IPFS node is that Fleek Network only caches the files, rather than store/pin them, which becomes unnecessary to do given this new setup (further explained below).
Making all Content Retrievable by CID from Any Origin
Moreover, Fleek Network stores a map of all uploaded content CIDs to their origin location(s) in perpetuity to ensure that content is always retrievable through IPFS/Fleek Network, even if it has fallen off IPFS, so long as a valid origin exists.
In other words, Fleek Network brings the features and benefits of IPFS that everyone loves and is using it for (content routing and addressing) to EVERY underlying storage layer, making IPFS storage layer-agnostic, and removes the need to store/pin with centralized pinning providers at the IPFS layer without sacrificing anything.
In addition to the decentralization and trustlessness benefits, this setup will also lead to better cost and performance related to IPFS usage and retrieval, given the nature of Fleek Network as a decentralized cache layer.
These characteristics improves and evolves on IPFS in many ways, but these are the six main improvements it generates:
- Fleek Network opens up IPFS to be storage agnostic.
- Fleek Network hardens the availability of IPFS content.
- Fleek Network makes IPFS more performant.
- Fleek Network makes for a more decentralized and sustainable gateway to IPFS.
- Existing IPFS gateways can leverage Fleek Network for further decentralization.
- Pinning providers can leverage Fleek Network for further decentralization.
Let’s break these down, and cover how Fleek Network achieves that.
1. Fleek Network Opens up IPFS to be Storage Agnostic
Since Fleek Network is built with IPFS primitives (content addressing and routing) and can interface with any underlying storage provider (web2 and web3), Fleek Network allows developers to build with IPFS without having to store / pin directly on it.
Instead, they can store on whatever underlying storage layer they want (Filecoin, Arweave, etc.), and since Fleek Network stores the CID/origin mapping in perpetuity, you can always access your files from the consistent content identifier (CID) at any point.
2. Fleek Network Hardens the Availability of IPFS Content
Unlike using an IPFS pinning provider where content can become unavailable if it falls off IPFS (or if the pinning provider goes out of business), content that falls off Fleek Network is always re-fetched and trustlessly brought back onto the network.
How? While IPFS can’t map to other underlying origins, Fleek Network keeps a map of CIDs to origin(s). That way, when a piece of content has fallen off the network’s cache and is requested, it is automatically fetched from its origin as defined in the mapping. Given a CID, Fleek Network will always return a file, if a valid origin exists. Long-term, that is more sustainable, trustless, and decentralized.
3. Fleek Network Makes IPFS More Performant
As an edge and content delivery network, it can be expected that straight out of the gate Fleek Network nodes will be more powerful (and faster) than traditional IPFS nodes and gateways.
This is by design. Fleek Network will have an 80% to 90% cache hit rate, meaning most of the time content won’t even touch IPFS but rather be served blazingly fast from the cache of the Fleek Network node that is geographically closest to the requester. Given all centralized pinning providers use CDNs in front of their IPFS nodes, it’s the same setup as today, the difference being Fleek Network is decentralized whereas Cloudflare (the most commonly used CDN in pinning) is not.
4. Fleek Network is a Better, more Decentralized, and more Sustainable Gateway to IPFS
IPFS gateways are really just a fancy way of saying a place where you can access files from IPFS – this makes Fleek Network a huge, decentralized gateway to performantly accessing IPFS content!
With Fleek Network, gateway node operators can set up one or multiple access points (ex. different URL’s) to the network, and since all Fleek Network nodes also serve as IPFS gateways, any of those url’s/access points can be used to access content on IPFS, even if it isn’t already cached on Fleek Network. Gateway node operators or other service providers could even offer a service to allow you to spin up your own public or private gateways on Fleek Network.
To add, all major public IPFS gateways are currently subsidized, meaning they are allowing people to use them to access IPFS content for free. However the organizations behind these gateways can’t just give away free bandwidth forever, and Fleek Network is a much needed long term sustainable solution to this impending issue. Evenprivate IPFS gateways are in their majority run by centralized infrastructure owned by corporate entities. Therefore, private gateways on Fleek Network will also be far superior, being run on top of decentralized infrastructure.
5. Existing IPFS Gateways could Leverage Fleek Network for Increased Performance and Decentralization
All current IPFS gateway operators already use a CDN in front of their IPFS infrastructure to increase performance and deliver content as fast as possible. The issue is they all use Cloudflare, which is centralized and prone to censorship. By replacing Cloudflare with Fleek Network, IPFS gateway operators will be more decentralized, censorship resistant, and also reduce costs without sacrificing performance.
6. Existing Pinning Providers could Leverage Fleek Network for better Decentralization and Censorship Resistance
Similar to the IPFS gateway operators mentioned above, all current IPFS pinning providers already use a CDN in front of their IPFS nodes to increase performance and deliver content as fast as possible, while also lowering costs (because it’s way cheaper to serve content via cache/CDN rather than storage).
By replacing Cloudflare with Fleek, current IPFS pinning providers will remove a huge central point of failure from their infrastructure (Cloudflare), making for a more decentralized and still performant IPFS pinning setup.
In addition to that, pinning providers could also opt to move away from their own centralized pinning/storage architecture. Currently they store all the IPFS files they pin on their own infra which is usually running on AWS or other cloud platforms. Instead, they could leverage Fleek Network to allow storage on any underlying storage layer, including decentralized web3 storage protocols like Filecoin, Arweave, etc. Making them more web3 aligned, and also reduce their own infrastructure costs as storing on those networks is way cheaper than storing on AWS.
We can’t wait to see Fleek Network move forward into its Devnet and Testnet phases, where it will show the impact of these improvements live! Sign up for updates on our website to not miss each new development.
As an IPFS-friendly network, leading the evolution of this technology makes for a terrific challenge to take on. There has not been major innovations behind these important primitives IPFS provides, nor a push to decentralize its offering - and it’s exciting to see Fleek Network as a force of change ready to address that.