January 18 20: 00, BlockMania AMA live on 43 to continue, the current event is organized by well-known technology community Comunion, BlockMania Co, the current theme of "consensus from the floor to see the chain block chain block application", the guests share Fu Xiannong to FileStorm founder, Tsinghua University graduate, received a master's degree in computer United States Georgia State University in the US IT industry for 20 years, has a wealth of experience in the industry. The main research interests include block chain, big data and artificial intelligence, business intelligence, and enterprise software architecture.
The following is the full review AMA
What block chain consensus is that?
Consensus mechanism is the soul of the block chain. Like a country's laws, to maintain the normal operation of the block chain of the world. The first block chain consensus is that the workload of proof (POW), workload and Ethernet Bitcoin Square are based on proof.
Now a popular trend is proof of interest (POS). Ethernet Square 2.0 will use POS consensus, as well as the latest public chains such as Tezos, Polkadot, Algorand and so on.
Proof of interest also include entrusted proof of interest (DPOS), on behalf of the public chain has EOS. DPOS consensus FileStorm now is the use of chain alliance will also PBFT, RBFT and so on consensus. Consensus is the soul of the block chain, is the basis of the value of the block chain.
Details See also:
https://www.jianshu.com/p/2f29f3909b11
Why do we need consensus?
Technically, because the block chain is decentralized. All participating nodes need to have a consensus in order to maintain a single chain. This option only chained, is a rule, we can be competitive to get, it can be drawn to vote.
This rule is consensus. So it can be said, the purpose is to determine the block chain consensus right out of the block.
The basic idea is to work with a proven force calculation to determine the right block. If you can first solve computational problems POW, you can block out.
At each point in time, we are computational problems, then at about the same time, we work out the problem, and then publish the results out. This is actually a voting process. The block is released the votes cast. Finally, the selected tickets, is calculated from the results of the fastest mining machine. As for node does not release the block, which is equivalent abstained.
Workload proved one of the biggest advantages is that the votes cast and the voting rights are binding. And once the vote is successful, even if I can not modify the contents of the voters vote. Because the general in front of the block, the node has packed a block, then the block fishes work to do to prove, cast out immediately after the finish. This time packed block has no way modified. So great effort to prove safety guarantee.
Drawback is that the delay is relatively high, because we have to deal pack after block, but also to complete a workload proved that this block can be called candidate blocks. Workload proved particularly time can not be set too short, or easy to fork.
So, at least from the packed blocks to the finish block workload proved this period of time, the transaction can be confirmed.
Another POW mechanism has been criticized point is particularly high energy consumption, environmental damage. Because the hashing power. So, to get the block chain support large-scale commercial, POW needs to change. So, there is a POS.
With POW not know how many participated in various ore mining machine, POS protocol must know their number and hold out the consensus of all participants of the amount, and then to allocate packaged and voting rights.
Therefore, based on the consensus of POS everyone how much of a chain of economic investment. POW in the inside, and voting rights are packaged together, but at the POS where the two can be separated.
There blocks before you know who has the right to vote. After get the right to vote, they can easily go to vote, which means that I can choose a larger strategy space, the entire game will become more complicated.
In general, a large space policy for security is not good, because the attacker aside more room for maneuver. This safety chain is challenging, because the nodes through multi-cast a vote or abstain from voting for evil.
But in the POS system, the nodes do pledge, if the participant has detected a violation of POS consensus of agreement, can make some of their punishment. DPOS commissioned proof of interest are POS variants.
Because decentralized mechanism, we need to elect Who is responsible for packing, who is responsible for voting. Sometimes the interests of efficiency, it will first select a relatively small committee, then in charge of the vote by them. Such a committee is usually chosen at random, to fairness and even frequent rotation.
Or you can also use proxy rights and interests of DPOS proof mechanism to select a relatively fixed charge of packaging and vote in the Committee. While some of this loss to the extent of decentralization, but can greatly improve efficiency.
FileStorm DPOS mechanism is used, help us promote our node platform, which is why our platform can develop so fast.
We can no longer explain in detail the chain block some of today's more popular consensus mechanism?
The first block chain consensus is that the workload proved that we are more familiar. BTC, ETH, and a lot of work out before 2017 are based on the vast majority of the public chain of proof.
The new public chain in most cases, POS, so, I'm here to introduce several POS consensus, mainly to understand their rationale and their differences.
1、Tendermint
This is the consensus of cross-platform Cosmos chain used. Getting to know the consensus can be used as a POS. Tendermint each release by the new block is not fixed nodes, and let all the other nodes through PBFT way to block vote verification. More than two-thirds of the votes to pass, and is the ultimate confirmation of the block, unlike the POW have to wait for several blocks. But the disadvantage is that if the vote pass, we should have been waiting for.
2、Polkadot
Boca is a chain of cross-platform, it uses a relay chain Nominated Proof of Stake, both "nominated proof of interest." Participants pledge to become the nominee by then elected from nominees verifier, the verifier then in charge of the block.
Number verifier can be controlled, democratic centralism, so that efficiency will be much higher. NPOS with DPOS like, are chosen few selected node out of the block, but DPOS elected are big chaebol, need to do a lot of the pledge. NPOS more democratic, grass-roots political participation easier, but also more frequent rotation verifier.
3, Algorand
As mentioned earlier, a POS nature of the problem is a block consensus by everyone the right to vote, so the block before we will know who is a block node, so vulnerable to attack; node can also choose to block out all kinds of packaging policy to maximize their own interests.
What if there is a mechanism to generate a random node, this randomness can not be easily predicted, but it is all very easy to verify, it would in this way can be used to block the election of a node.
Algorand to find a way to generate such a random node. This method is actually just like POW way very close. Everyone can be packaged submitting candidate blocks, but in the end not to see who is the fastest to calculate the title, but by a true random number, select only the next block.
4、Casper
Casper是以太坊 2.0 将要采用的共识方式，Casper有了权益质押，有了投票机制，但现在还是一个 POW 和 POS 的混合体，将来要慢慢发展成一个完全的 POS 协议。
因为要对现有的基于 POW 的以太坊进行升级，它考虑更多的是链的可用性而非最终一致性。Casper出块还是会通过 POW，然后再由 POS 节点投票。
投票的节点有点像是在赌博，可以随便投任何块，但是如果你没有投中最后被选中的块，会受到惩罚，所以最后大家都会趋向投最长的链上的一个块。
5、FileCoin
这是我们分布式存储行业里的一个明星项目，他采用了一种新的共识机制叫预期共识。这可以理解成一种新的 POW 或者 POS 模式，说它是 POW，因为他把算力挖矿改成了存储挖矿，说他是 POS 是因为它给投票权的理由不是看 Staking 里的通证数，而是看你的有效存储量。
其实每个项目的共识都很复杂，这里只是蜻蜓点水一样的讲一下。
从前面的分享中，大家可以看到，其实 POW，POS，包括 DPOS 都不完美，未来 FileStorm 将继续研究一种理论上无懈可击，又具有实用性的区块链共识。
FileStorm 采用的 DPOS 共识，前期对我们的项目启动帮助很大。但是未来我们希望能让我们的项目走向更加去中心化的公链共识。同时实现多链架构，更好的支持更多的存储应用。
从区块链应用落地看区块链共识，能否谈一下落地应用和共识的关系？
刚才一直谈共识，为什么需要共识？因为区块链改变的是人类的生产关系。以后没有老板和员工了，不存在谁为谁打工，而是大家为了共同的目标一起奋斗。
但是每个人的诉求是不一样的，为一件事愿意做的付出也是不一样的，如果希望通过区块链的激励机制来实现利益的公平分配，大家就需要对区块链高度认同，这个认同就是共识。
POW 通过算力投入实现共识，POS 通过资金投入巩固共识，共识的价值通过通证来体现。共识越强，通证的价值就越高。但是，如果这个通证不能真正实现应用，那这个通证就是空气币，就迟早要归零。
比特币因为是世界上第一个区块链的实现，共识强大，所以虽然它还没有实现中本聪赋予它的成为点对点的电子支付系统这个使命，它依然有价值。它是加密货币的黄金。
但是以太坊的价值来自于它实现了智能合约。这个智能合约让开发基于区块链的应用成为可能。代币发行，去中心化金融，都是它的应用。
FileStorm 的价值来自于发行的通证可以用来做存储。用户用通证购买存储，矿工通过提供存储服务来赚取通证，这就形成了一个经济闭环。这个经济闭环就能体现区块链的价值。很遗憾的是，当今实现经济闭环的区块链应用很少，FileStorm 是为数不多的一个。
但是我们要乐观的看问题。整个区块链行业还很新。基于区块链的应用占有的市场份额还很小，前途无量。使用我们存储的用户还远远不够，我们的业务未来会成指数级增长。我们的通证价值也会不可估量。如何让更多的用户知道我们的应用，就需要更多的人帮我们去布道，去加强大家对 FileStorm 的共识。这也是每一个区块链应用必须走的路。
如何通过一个强大的共识来实现落地应用？FileStorm 是怎么做的？
如何打造一个强大的共识，我们认为有两个重要点，一个是技术，一个是治理。
FileStorm 在共识机制上追求创新，就是从技术上优化我们的共识基础。前面我们提到我们要打造我们民族自主创新的公链，走向世界。这不是一句口号，而是我们真正要在技术上跟国际知名的公链项目较劲。未来大家一定会看到我们逐步推出的，让大家眼睛一亮的核心技术。
另外，区块链项目一定要开源。这才是公链项目正确的打开方式。FileStorm有几十个技术创新，github发布数百万行代码，也没有申请一个专利。
优秀的区块链项目都是开源的。
越开源越安全：比特币因为开源，越来越难因为代码漏洞被攻击。越开源生态越大：以太坊开源，无数项目抄袭，但它得到的是全球最多的区块链开发者为他服务。
在治理方面，未来区块链在持续开发中的进化和升级能力，将成为它能否在它所控制的利益竞争中生存下来的关键因素。
所以，区块链治理的根基，来自于这样一个事实：一个区块链系统有很多的涉众，矿工，矿场主/节点，投资人，用户和开发者。他们通过与区块链交互获得价值，成为利益相关者。
利益相关者一定是要维护自身利益的，而他们之间的利益是有差异的，从而他们一定会要让区块链为他们服务。
因此，如果要让整个区块链系统长期生存下去，就需要有相关的流程，来消除这些差异。创造和执行这个流程，就是治理。
我了解到 Comunion 就是看重了它在社群治理和社群协助这个领域的努力，我们的理想是一样的，区块链就应该这样的方式发展。
有了强大的技术和强大的社区，应用落地就不是项目方一个人的事了，所有的参与者，都会为项目添砖加瓦，寻找落地机会。
区块链发展到今天，有很多项目追求的东西太好高骛远，谈远景，谈高大上，但是不能在可预见的未来实现。这样的项目就需要超乎想象的强大共识，或者国家和大投资机构的信任背书，才又机会存活下去。如果不接地气，光吹牛，那很快就会玩完。
我们选择存储，因为这是跟每一个人的生活都息息相关的事情，每个人都可以是我们的用户，每个人都能为我们做项目推广。所以，我们选择的赛道，非常接地气，相当容易落地。
能否谈一下FileStorm 存储技术的特点？
首先，FileStorm 结合自身的技术特点，提出如下几条价值观念，推出了一个「数据存储的核心价值观」。我们还在跟整个行业探讨，现在提出的数据存储的核心价值观就是 保护隐私，数据确权，鼓励实用，技术开源。
数据存储的核心价值观反映的是数据存储领域的丰富内涵和实践要求，是人类对数据的创造和使用诉求的高度凝练和集中表达。
我们的创新技术如下：保护隐私就是要做到数据加密。数据加密大家很了解，也有很多现成的技术实现。但是高效的对加密数据进行搜索和计算却很难。我们跟大学和研究机构的顶级专家合作，共同创新同态加密和零知识验证技术，让加密数据搜索计算和普通数据一样。
有了让加密数据搜索，我们就可以把目标剑指分布式存储上的谷歌和百度。数据确权就是要给数据做时间戳，要做内容智能辨识，这个需要很强的区块链技术和大数据技术。这两个板块正是 FileStorm 的强项。（也是web3.0的技术核心）
另外，数据确权也不是一个应用能实现的，而是好多不同的应用。FileStorm 未来的多链生态，就是为这个目的打造。
鼓励实用就是要让数据动起来。FileStorm 的检索挖矿，按数据使用率给出奖励的共识算法，鼓励数据大量复制，大量分发。冷数据可以去重存储。但只有流动的活数据，才更实用，才能创造更多的价值。
技术开源前面已经提到，不开源，口里喊着去中心化，走的还是中心化的路子，是没有前途的！
所以，FileStorm 愿与全球分布式存储行业的项目一起，实现整个数据行业的价值目标，让数据更好的为人类服务，让人类真正成为自己数据的主人。
Q&A
Q1：共识的前提是需要有共识者，一般是如何寻找共识者呢吗？非公链项目如何设计共识呢？
您这里有两个问题：
公链如何寻找共识者。其实1CO是一个公链寻找共识者的好方法，因为一开始就从经济上把共识者绑定。ETH就是一个非常成功的案例。可惜后来被玩坏了。DPOS 其实也是一个很好的方法，可以早期就找到金主做盟友一起推广。但是缺点是盟友可能成为利益冲突者。权利太大，影响共识发展。
非公链项目如何设计共识：这个用联盟链的方式可以实现。其实比公链简单。
Q2：能评论一下POC共识吗？
您说的 POC 是指的 proof of capacity 硬盘存储空间挖矿对吧？这个跟我们存储挖矿行业很接近，因为用的是硬盘空间做共识。
跟比特币用算力挖矿不一样的是。POC 提前把哈希运算做好，存在硬盘里，像存了一堆彩票。然后到点抽奖，抽中了谁，谁就出块。
Q3：您是怎么看待集权与效率，投票与共识的呢？
这个问题非常好。也是我们常常讨论的。集权提高效率，但是牺牲去中心化。投票民主，但是效率降低。
我觉得每个项目要根据具体情况选择共识。看注重的是效率，还是民主。
Not every project is the pursuit of high TPS, this is the case, select the slower the block chain is also quite good. FileStorm want to do more than the ecological chain, it is to see every project is different, in the future we hope to support all types of applications based on consensus.
Q4: Do you think consensus and efficiency can coexist it? If so, then what way?
Now is the coexistence of it. Just the way the game is based on mutual co-existence. High efficiency league chain. If this consensus can be recognized by the League chain is enough!
As fundamentalist idea block chain believer, in fact, I had more public support chain. Until I see Google's Libra, a huge coalition chain platform, as long as the support of many people, as a very strong consensus.
Q5: Now the underlying component or protocol in addition to the IPSF there are other similar tile-based distributed storage chain it?
Many underlying protocol, not necessarily IPFS. We have done a lot of renovation on the basis FileStorm also IPFS, which is FileStorm latest data dissemination protocol layer architecture.
