I first came across IPFS (Interplanetary File System) in 2016. It wasn’t long before it convinced me that it would be the ideal data storage system in a blockchain-powered future. Like blockchain, it perfectly embodies the spirit of decentralization. Together, IPFS and blockchain will resolve the issues of computation and storage in an age of distributed computing.
To understand what I’m getting at, I should first clarify the relationship between IPFS and FileCoin. IPFS is a protocol. It is based on BitTorrent, which we know is a relatively mature technology. IPFS introduces DAGs (directed acyclic graphs), to adapt BitTorrent into a means of organizing and storing files. Meanwhile, FileCoin is an official currency developed for the IPFS protocol, currently in development.
IPFS is stable, awesome, but also imperfect
As I mentioned, IPFS is based on a mature technology known as BitTorrent. It doesn’t drastically alter BitTorrent’s core technology, and in theory is compatible with existing BitTorrent software (although it hasn’t implemented that in practice).
IPFS’s biggest innovation is using DAGs to organize data into a collection of seed files, thus building an organization system out of file systems. The fact that the hash paths that IPFS uses have an infinitely large namespace means that hypothetically, everyone in the world can put all of their files in the same file system. This gives rise to the “interplanetary” aspect of IPFS’s namesake.
IPFS’s second biggest innovation is the idea of DWeb (distributed web). Our internet today is built with the HTTP protocol. HTTP is a centralized data delivery framework. This leads to lots of wasted bandwidth, and an overreliance on major global networks. This also makes it easier for some countries to exert more control over their internet, and for some hacker groups and individuals to achieve their goals (cough DDoS cough). In comparison, IPFS Web is a P2P node network with no designated server.
Caveats to IPFS
- IPFS’s ecosystem lacks an incentive to ensure operational nodes
IPFS itself is just a technology protocol. Without an appropriate ecosystem and incentive structure, the technology is moot and cannot realize its true potential.
- DWeb is hard to actualize
As of today, we don’t have a promise that Web browsers will support IPFS. There has been discussion about using IPFS_API to connect to designated nodes, but all the advantages put forth so far are incomplete and self-inconsistent. More importantly, this proposed workaround by connecting to a designated node reduces the whole system to a centralized system once again.
The potential for innovation and imagination with IPFS has garnered much attention around the globe, with many people trying to help it overcome its biggest obstacles. The Hero Node project and IPFS’s official FileCoin project are just two such examples. At this point, let me remind us that IPFS itself is just a technology protocol, which means the protocol poses no risk of failure. However, I will discuss how FileCoin may fail.
What you may have misunderstood about PoSt/PoRep consensus protocols
When talking about PoRep (Proof-of-Replication) and PoSt (Proof-of-Spacetime), the two main consensus protocols for FileCoin, I’d still like to see some comprehensive papers that fully examine their every benefit and risk. As of now I have yet to see any such comprehensive writeup.
Let’s first consider PoRep. It is a form of PoS (Proof-of-Stake). This means it cannot be a network-wide authentication method, as a network-wide authentication method requires everyone in the network own the file. In reality, this is unfeasible or economically untenable.
According to FileCoin’s white paper, PoRep is an market, which means you need to have customers paying to store their files, in order for mining to happen. Conversely, suppose everyone rushed in excitedly to mine FileCoin — would that actually benefit anyone?
Now let’s consider PoSt. Slick name branding aside, simply put it means that you pay to store data, and you pay to access data. So again, it is an market. To say the 90% of its currency is reserved for mining is equivalent to saying the 90% of its currency is waiting to be sold. This is different from how mining Bitcoin or Ethereum works. I suspect, even guarantee, that mining in such a protocol won’t generate any returns.
An overly complex structure
FileCoin’s consensus algorithm is highly complicated. I spent many moons studying it and still struggled to get to its bottom. Perhaps it is because its architects instilled in it their vast ambition, or maybe it is because the market placed it in high regard. Either way, FileCoin can’t shed its sophisticated, complex image. However, advanced does not have to mean complex. Oftentimes, the simpler things are the ones that are quickly accepted by the market.
The Bitcoin whitepaper looks simple on the surface. The encryption algorithm enables immutability, PoW (Proof-of-Work) protocol ensures only a single ledger exists, and UTXO allows the ledger to be authenticated.
FileCoin attempts to build a bridge with every blockchain, to satisfy the market’s vast fantasy and imagination at the word “blockchain”. To say this is ambitious is an understatement, but also hints at a fatal flaw from the start. Cross-chain technology is a very hot technical area these days, but I think that just reflects a hype from venture capital. From a technical standpoint, this technology is nigh impossible.
Overly complicated mining design and imbalanced incentives make an inefficient system
Aside from the technical complexities, FileCoin’s ecosystem is also exponentially more complex than other blockchain projects. Miners are split into mining the replication market and the retrieval market. Add consumers and investors to the picture, self-regulating this ecosystem may prove to be difficult or self-consistent. Additionally, consumer perceptions may lead to legal disputes in the event of lost data.
FileCoin does not provide a free storage protocol, which means a high barrier to user entry. FileCoin also does not guarantee eternal storage. Once FIL (FileCoin) is fully spent, data loss will begin. This is inconsistent with the idea of eternal storage shepherded by blockchain. The problem arises from the fact that FileCoin is in itself a market, and not a community with a balanced rewards structure. Most blockchain evangelists lose faith in FileCoin once they realize this fact.
Targeted towards a niche market
When we first encounter IPFS and DWeb, we may think FileCoin can be used to power the distribution of Web data, and the sharing of streaming media. But once you take a closer look at the white paper, you realize that these are not FileCoin’s market goals. It charges consumers for retrieving data, while Web users and streaming media users won’t want to pay extra for their bandwidth usage. FileCoin is only tailored for the private cloud storage market, which is far too niche of a market.
How is Hero Node any different?
Here are Hero Node’s responses to the two major issues hampering IPFS:
Issue 1: IPFS’s ecosystem lacks an incentive to ensure operational nodes
Hero Node is set up so that users are encouraged to operate Hero Node nodes (which have built-in IPFS nodes), thus providing public storage services for applications, and earning automatically created tokens from the system. We will shortly discuss why these tokens are automatically generated and not paid by the application developers.
Application developers can publish their application onto the node network for free. Only when they want to make their application gain broader distribution, do they have to lock in more tokens to smart contracts. The more commercially successful an application is, the more token it needs to have locked in. This gives us a constructive incentive structure, where everyone in the community has their interests aligned.
Issue 2: DWeb is hard to actualize
Hero Node’s applications are distributed applications (DApps), and not DWeb, so it has no browser constraints. This enables IPFS to truly replace HTTP. While it is true that Web is a very mainstream application format, Hero Node itself is also a node network. So even in a Web environment, the vast number of Hero Node nodes connected via IPFS_API ensures that we have full decentralization. The only caveat is benefits of CDN are not fully brought out.
What’s Hero Node’s answer to FileCoin’s problems?
- Hero Node breaks its technology into three steps, and avoids setting overly lofty goals that are unattainable in the short run
Step 1 — Get the ball rolling now
Even without a currency system in place to create incentives, early investors and crowdfunding efforts will provide resources to support the deployment of official nodes. This enables Hero Node’s decentralized DApps to quickly start running on the network
Step 2 — Node-based rewards structure
External miners can join the action by operating their own nodes, but at the moment the currency used is still ERC20, the Ethereum token standard. Smart contracts will be used to reward node contributors.
Step 3 — Engage the community
We are in researching consensus algorithms based on Proof-of-Work and Proof-of-Capacity, which will be used to validate the quality of public sync services and the efficiency of IPFS storage provided by nodes. Our HNC token reserved for mining is directly issued to miners via consensus, and not paid out by the application developer.
- Hero Node seeks to simplify the framework design of technology, and offer a flexible value proposition
In Hero Node, the number of tokens miners receive is regularly decreasing, and independent of the token’s market value, market sentiment, or any other external factors. Application developers don’t need to spend tokens to use network resources; they just need to lock in tokens to increase the distribution of applications that bear their signature. This increase of distribution is also probabilistically determined by programs. On a macro level, the probabilistic aspect of these programs injects an element of fairness and flexibility into the system.
- Hero Node’s market is aimed at all internet application developers
Developers can develop any application, and Hero Node’s potential target industries include cloud computing, cloud storage, CDN, app stores, advertising, and more.
Moreover, Hero Node integrates IPFS across many different public blockchains, to create a fully decentralized DWeb. It then layers on cross-platform development technology to turn DWeb into DApp, fully bringing out IPFS’s advantages and greatly improving user experience. Finally, Hero Node uses the power of node networks and the general community to make sure DApps stay online forever.
Hero Node is introducing an authentic, easy-to-use, eternal DApp
Forget FileCoin, because Hero Node is going to be the Player One of IPFS
Wait, you’re still here? Man, that was some really dense stuff, kudos on making it to the end. Here’s my WeChat contact (gpliuwind), shoot me a message, maybe I’ll throw out some coins :D
Guoping Liu
Blockchain technology and applications expert, one of Bitcoin’s first miners, founding developer of the Hero mobile cross-platform framework, president of the Hero Council. Formerly @ Wind, The9, and Dianrong, where he was heavily involved in the development of Dianrong’s blockchain applications.
Previously spent many years developing the Hero framework and blockchain applications. In 2017, meshed the two together to create Hero Node, aimed at making distributed applications easy to develop for everyone. Staunch believer in a better future with distributed applications.
In an open world, Hero Node welcomes any and all feedback. Especially helpful feedback will be rewarded!
---------------------------------END------------------------------------
Talk to us:
Email: feedback@heronode.io
Twitter: Hero Node