What is hyperledger - Discussion


#1

Welcome to the discussion about this lecture. Here you can ask questions or post feedback about this specific lecture.


#2

Hello Filip,
A quiz was mentioned also in this sections video. I do not see it. Thanks.


#3

Thanks for letting me know. I have published all the quizzes now.


#4

Ah, Ok. Thank you. Also, please disregard the other email regarding same situation but on a different topic.

Have a super day!


#5

Great! Yeah I made a mistake earlier. But now everything should be right. Sorry for the delay.


#6

Sorry not I had to link this to a previous email/topic but I couldn’t figure out your address to compose a new email.

Hi Filip,

This inquiry is not for the mini quizzes on each of your lectures as we have recently corresponded about (i don’t like using the word lecture, it sounds so monotonous). Let’s re-word it to…each of your videos. Anyway…this is referring to the Hyperledger Quiz at the end of the “What is Hyperledger” section.

After all that my point is…I cannot access it.

Please let me know if it is an issue on my end, or if you possibly haven’t unlocked it.

Thank you for all you do.


#7

The end quiz is now removed. Instead of having one quiz at the end I created smaller quizzes after each video. It corresponds better to what I actually say in the video. So there is no issue on your part here.


#8

I’m just learning about permissioned blockchains like Hyperledger, but I’m feeling a bit confused. As far as I’ve understood, the big thing about blockchain is to have trust “outsourced” to a system. For “the first time in history” we’re having computer systems that can contain immutable data and because you can’t change it without following the preset rules, you can rely on the information (assuming there is a large enough, decentralized network to avoid 51% attack, etc. etc.).

From what I hear about Hyperledger (or rather its implementations), on the highest level, you don’t have different parties owning different parts of the network, i.e. no decentralization. How can you trust it? How can it be a “blockchain”? Sure, I can see that it contains a blockchain data structure and perhaps put more strict rules for each (non-admin) participant. This can be a good way to force everyone in line and increase abilities to reason about the data. But how does it differ from the distributed computer systems we’ve “always had” (albeit more modularized/pluggable) say an event-sourced, clustered database coupled with an identity service.

In short why are permissioned blockchains even compared to public, permissionless blockchains when they are ultimately centrally controlled? From my understanding so far they don’t exhibit any “novel, trustless features” they just seem to mimic them.

(Disclaimer, I’m not trying to spread any “FUD”, I’m asking sincerely…)


#9

Very good point. This all comes down to implementation. Do you have one company governing the entire blockchain or is it controlled by for example a consortium of companies. What type of enforcement rules are put in place in order to establish the more or less immutable characteristics. This is all up to the company. I would say there can be good implementations of private blockchains but also a lot of bad implementations of blockchain. In the bad examples they are just a more expensive and slow version of a traditional database.


#10

All right, thanks for the answer! It makes sense there is a scale of “trustlesness”. But I think I read that private blockchains don’t contain any cryptocurrencies. If there is a consortium of companies collaborating, isn’t it still hard to achieve “decentralized trust” without any currency to mediate and create incentives between equal parties? Perhaps my question is too vague and boils down to philosophy or definition of terms. Thinking freely now, I guess the governance of a democracy where you have “separation of power” like “trias politica” is an example where none has full power, still without any market. But it feels very complex to design properly…

It was interesting to hear about concrete examples, like the one with Maersk, IBM etc. Does Hyperledger themselves have any goal to interconnect different Hyperledger blockchain projects or is Hyperledger only focused on providing the tools to build blockchains?


#11

as per video it states that Intermediate CA can create certificate but signing of the same will be done by Root CA only and later when Root CA revokes its sign the Intermediate CA gets detached.

But in Quiz 3 1.What is a Intermediate Certificate Authority?

answer for which currently is
An authority that can sign certificates on the behalf of the RCA

Please share some thought on this, what i think answer would be

A certificate holder that is allowed to issue new certificates.


#12

Yeah it’s definitely a difficult and complex thing to create a private blockchain with the right incentives and a governance policy that actually works. I think you would have to align incentives outside of the chain/currency itself. The incentive should probably be tied to ones reputation in the industry or the reputation of your product or service.

It’s also very difficult to know how IBM and Maersk have designed their system, because it’s not open source… But I haven’t heard of any efforts from hyperledger like that, as far as I know hyperledger just provides the tools. But I may be wrong.


#13

Yeah I understand if that can be confusing. Those two statements don’t have to be mutually exclusive.

The Root CA will sign the ICA certificate and through that it has “delegated” some of it’s power to the ICA. From there the ICA can sign other peoples certificates and those signatures can be traced up to the Root CA, and therefore they become trusted.

Do you understand my explanation? In the end it’s just different terminology.


#14

Right, thank you for the answers, they make sense. I wish the incentive structures were more clearly stated by different projects, but of course, they may be delicate to share.


#15

I’m a little confused about the peers. The diagram in the lecture shows a 1 peer/node block chain (I presume). I’m trying to envision such a system in a multi-node block chain. You mentioned the use case where there’s many many applications pushing many many proposals and each node is processing the chain code, registering the transaction in the ledger and then forwarding a call to the ordering service which presumably finalizes the entry in the ledger. That all makes sense to me in a single node blockchain.

How are the ledger transactions broadcast in a multi-node system. I remember reading about parallel execution, so it seems it would have to happen at the earliest onset. As in, when the first node gets the request, it’ll broadcast it to the other nodes, they all process the request . . .I presume whoever finishes first gets to call the ordering service, which then passes the final sequence to all of the node’s ledgers. If one of the nodes ledgers is out of date, their ledger will get corrected. Am I close to understanding this concept??


#16

Ha. I guess I should have watched the “Channels” lecture before asking my question. I think I’m good.


#17

Helo Filip,

In the quiz, there was the question: where is the world’s state stored?
and the right answer was: in each peer.

I thought, that the world’s state is a pluggable database (at the moment LevelDB or CouchDB) in the Hyperledger Network. Is there such a database for each channel? if not how is confidentiality assured?
Thanks in advance
Michael


#18

Correct. It is stored in each peer, just like the blockchain ledger. It is pluggable in the sense that hyperledger supports a variety of different databases, you can use whichever fits your use case.

As for channels it’s the same as for the blockchain ledger itself, each peer holds that database. I hope that answered your question, otherwise let me know.


#19

@filip , this forum page is linked to the “Channels & Confidentiality” course page. I think I’ve spotted a few “singular vs. plural” grammar errors in quiz headers. I might be mistaken as I am not a native speaker but you’d better double checked and proof-read the content anyway.

Why do each channel have its own MSP?

Why does each channel have its own MSP?

Does all channels have it’s own ledger?

Do all channels have their own ledgers?
Does each channel have it’s own ledger?

P.S. No offense. I’m only letting you know as I believe you’ll take the steps to improve your course content.


#20

@etsit , I believe hyperledger is used in order to guarantee transparency and avoid cheating when several companies interact with each other.

From what I hear about Hyperledger (or rather its implementations), on the highest level, you don’t have different parties owning different parts of the network, i.e. no decentralization. How can you trust it?

For example, you have some deals. And as different phases of the deal implementation take place, the corresponding records are logged to the blockchain. Typically they are signed by several parties involved with the deal. Think of it as the company representatives signing off the documents. Since the blockchain is immutable and the company representatives have left their verifiable signatures, the blockchain data can be used as a source of truth during auditing or during disputes between the companies.

But how does it differ from the distributed computer systems we’ve “always had” (albeit more modularized/pluggable) say an event-sourced, clustered database coupled with an identity service.

The key benefits are

  1. The blockchain is an “append only database” which is almost impossible to tamper with. As opposed to traditional DBs having the DELETE and REPLACE instructions.
  2. In case of traditional DBs each company would have to maintain its own DB replica which can be potentially tampered with in order to cheat on other companies. With blockchain you maintain a node but it is secure and can be trusted (which is guaranteed by cryptography and other technical aspects of the blockchain technology)

In short why are permissioned blockchains even compared to public, permissionless blockchains when they are ultimately centrally controlled?

To emphasize that they have completely different use-cases, I guess. Many newcomers might be confused on the difference between these, so the mentors pay this much attention to the differences.

From my understanding so far they don’t exhibit any “novel, trustless features” they just seem to mimic them.

The idea is to make accounting and dealing more transparent and secure by the means of technology. Which allows to save the money paid to the employees whose job is to ensure the other company is not cheating.