Editor's Note: This article comes fromVernacular Blockchain (ID: hellobtc), Author: Five Fireball Leader, reprinted by Odaily with authorization.
Editor's Note: This article comes from
Vernacular Blockchain (ID: hellobtc)
Vernacular Blockchain (ID: hellobtc)
……
, Author: Five Fireball Leader, reprinted by Odaily with authorization.
Since the development of the blockchain 10 years ago, we have initially realized several categories of "decentralization".
Ethereum, which realizes the decentralization of computing;Why is Filecoin, which has been delayed, so full of expectations? | An article about IPFSHowever, it was agreed to go online in 2018, and it took two years to skip tickets. Fortunately, at the end of last year, it finally entered the testnet stage, which made many people see the dawn. Now, it is only more than 3 months before the official mainnet launch. up.
For the basic introduction of IPFS, you can read an article before the vernacular "
Why is Filecoin, which has been delayed, so full of expectations? | An article about IPFS
", today, when the dawn is coming, let's look at IPFS or Filecoin from another angle.
secondary title
Decentralized storage, is it necessary?
In the territory of each "blockchain" category, one of the first questions to ask is-is this thing necessary to be decentralized?
Many old leeks still vaguely remember the madness of the coin issuance tide in 17 and 18. At that time, the project party was like looking for a nail with a hammer, and it was like holding the "blockchain" above the sword. To centralize one, as long as you can think of, such as Facebook, WeChat, taxi, food delivery, Github, Airbnb... All centralized Internet applications can basically find one or even several corresponding blockchain projects. Moreover, the retail investors at that time still followed this routine, desperately looking for quotas everywhere, "Come to me, I will give you the money!" Holding a ticket in their hand, they waved their hands at a seemingly high-ranking project party.
Everyone knows the ending, the tide receded, and a large number of naked swimmers showed their true colors. This frenzy of "decentralizing everything" finally died down.
So the question is, is decentralized storage really necessary?
This is really a question of the benevolent see benevolence and the wise see wisdom.
In addition to the decentralization of Bitcoin and Ethereum, no one is arguing, whether the tools that serve Ethereum, such as oracle machines, should be decentralized is still controversial.
For example, some people think that the oracle problem in the real world is easy to solve. For example, if you want to know the results of the game, you can go to the ESPN official website (AOL TV network) to watch it. Do you not believe the ESPN score? No, everyone believes in it. ESPN also has enough motivation to do things honestly. It doesn’t need all kinds of oracles. Even if it serves Ethereum, why use Link? Direct API access to ESPN is not enough?
But some people say, what if ESPN fails to publish to the blockchain correctly? The probability of single point of failure and error is always high, including ESPN staff’s operating errors, hacker attacks, etc., and more Isn’t it a good practice to choose an oracle from multiple sources of information, and just use one of ESPN’s sources?
Decentralized storage also faces a similar "dilemma".
On the one hand, centralized storage currently has the best combination of efficiency and cost. In terms of cost, the number of copies of Amazon S3 is definitely much lower than that of decentralized storage like Filecoin. Generally speaking, the cost is either lower than or equal to that of decentralized storage. In terms of reliability, S3 is already 99.999999999%, basically Other than World War II or the sudden disappearance of Amazon as a whole, there are not many scenarios where you need to worry about data stored on S3.
What about data transmission? Another big advantage of decentralized storage is transmission. If 10,000 Shanghai users download the same movie on a Beijing server, the same duplicate data will have to be transmitted 10,000 times from Beijing to Shanghai, causing a lot of trouble. The transmission pressure of the backbone network is huge, and the user experience is not good. BT was popular in this situation. For a movie, everyone automatically finds a movie that is close. Take a little here, take a little there, and you can make it together. Spell out a movie. The download speed is fast, and it saves so much bandwidth for the backbone network.
Of course, the cost of CDN's computer room and bandwidth are not cheap, and the centralization attribute still exists. However, with CDN, CDN+copyright has basically killed BT downloads.
So what are the advantages of decentralized storage?
You may think it is a classic decentralization rhetoric, such as centralization to save and store these data in large data islands, they are easily sold, tampered with, extracted or otherwise abused, such as Facebook incidents, data ownership It should be personal Blah Blah…
However, the author thinks that there may be only one real advantage, that is-theoretically, it is possible to realize a file system that will never disappear. That alone makes it alluring. On a larger level, decentralized storage provides a solid foundation for the next stage of "information freedom" for human beings.
secondary title
Advantages of IPFS/Filecoin over competing products
The concept of the Ethereum global computer was proposed in 2014, and in the past 6 years, the progress has been slow and somewhat disappointing.
However, compared with decentralized storage, the development of the underlying smart contract platform is much better. At least we have dozens of competitive public chains such as EOS, with sharding, side chains, and cross-chains. Technology on the Internet, there are thousands of projects whether anyone uses it or not... oh, yes! We also have DeFi!
Decentralized storage is much worse. For example, Siacoin and Storj were established in 2015 and 2017 respectively. Maidsafe was established earlier in 2006. Well, yes, it is earlier than Bitcoin...
But now, after so many years, is there one that can be played, or how many people can really use it?
The answer is no.
Looking back at the public chain, or the smart contract platform, do you think it is not so bad? How else can we say that happiness comes from comparison...
Therefore, as the last hope of decentralized storage, it is almost the only hope. What IPFS+Filecoin carries is the "heaven will send a big responsibility to everyone" of the entire track.
So how is IPFS, or Filecoin, better than other storage companies?
1. It has a grander vision
The decentralized storage of other companies is more of a decentralized blockchain version of the current centralized cloud storage.
IPFS is aimed at lower-level things. It wants to fundamentally change the current Internet data distribution mechanism. The target is not only storage, but data positioning itself. Corresponding to current Internet data, PK is HTTP, the cornerstone of the contemporary Internet. general agreement. And the source of its technology itself is not passive water, but the DHT hash addressing technology used by BT downloads that were popular all over the world.
2. It's more famous and richer
In 2017, FIlecoin raised $257 million in less than an hour, and became popular overnight, breaking the ICO record at the time.
The investment capital behind it is all top institutions in the industry - Sequoia, Winklevoss Brothers Fund, Y Combinator, DCG, USV...
It may also be the only project in the industry that has launched futures on 8 exchanges without going online.
3. Its consensus mechanism is more complex and safer
Filecoin's 14-year white paper came out, and it has only reached the testnet stage now, and its technical complexity is also evident.
In order to prevent sybil attacks, external source attacks, and generation attacks, Proofs-of-Replication is designed.
In order to ensure that miners abide by the rules and effectively store user data, Proofs-of-Spacetime is designed.
At the same time, the entire protocol stack is also very complicated:
4. It already has the support of many industries and even ecology outside the industry
At the recent ETHDenver summit, 24 of the 52 projects submitted to ETH Denver were built using IPFS;
The underlying storage system of EOS uses IPFS;
Last year, Cloudflare and IPFS announced cooperation (Cloudflare is one of the world's largest CDN and DNS service providers);
This year, Netflix and IPFS announced cooperation;
A few days ago, Vitalik announced that ethereum.org, the official website of Ethereum, is now on the IPFS network;
There are many, many more...
secondary title
Where is the development? Is there anything to pay attention to?
This may be the question you are most concerned about. After all, tickets have been skipped for nearly two years. Is it expected to go online this year?
The approximate roadmap for this year is as follows:
The first phase of the testnet (November 19-April 20)
Testnet Phase 2 (April 20 to June 20)
Mainnet (July-August 2020) (recently delayed)
We are currently here:
Current status of the testnet:
7,At present, the two test versions of Testnet/2 and Testnet/3 are running at the same time. The Testnet/2 phase is coming to an end, and the Testnet/3 phase has been launched and gradually stabilized.
At the same time, recently, Filecoin officially released a "PB-level miner test" proposal on GitHub, inviting miners from all over the world to participate in the second phase of the Filecoin testnet test.
Is there any hope that it will be implemented according to the roadmap?
Personally, I feel that it is unlikely that there will be a big delay. After all, when it comes to the testnet stage, it is almost the last step
Can it be launched on the mainnet as scheduled in August? I feel a little overwhelmed, and if it drags on for a month and a half, I won't be surprised.
The reason is also very simple. It has been delayed for two years. If you look at last year’s roadmap, the time is several months earlier than now. Then the roadmap in a few months will suddenly be delayed by two months. Surprised, at the AMA held last month, the team leader stated that "the time published on the Gantt chart is the team's best estimate, and is currently very confident about this time, and there should not be any further delays"
Well, you see, "best estimates", in general, are just not likely to meet deadlines...
