Topic: "Blockchain-as-a-Service" does it worth to use? - page 2. (Read 956 times)

an append only database in a central store means everyone has control and access of a single file,
imagine 100,000 people all having access to your computer. you dont know who really just hacked it and deleted the file, you dont know who just scrubbed the log files and you dont know who could do it again.
also someone can then replace it with a different copy that they pre-edited/created before popping it into your computer.
so you employ IT guys, you get auditors in, you even employ HR guys to vet people that access it. security guys that approve or revoke access to people.

an append only database distributed across multiple systems
means everyones copy just adds new entries but has no hash or chaining function to check previous entries are not edited easily
meaning it requires more labour and time and money to do stuff

but blockchains without any labour self manages and audits all the systems and keeps the database robust. think about it have you ever needed to manually request your node to verify stuff.. have you had to manually ban mischievous nodes. or has the system been self reliant with no labour overhead/costs

imagine the international domain names registry.. different countries having a node.. it not public meaning users cant change entries they have to buy domains from domain services... and no single country can edit it,

then do some research on bitdns blockchain. its been around for 8 years

---

edit (gmaxwell has gone on a censorship power trip)
to answer points in posts below

ok take the UTXO set and imagine that it is just account balance,
imagine its backed up only once a day. thats 3.5GB a day.
imagine having to keep the backups for 10 years. thats 12 terrabytes
(you should already by now se the benefit. 250gb vs 12terrabyte)

also by only holding a 'current balance' log once a day. its not logging where thee funds came from, and its not logging all the movements during the day, its just showing the balance at a specific time of say midnight each night.

12 terrabytes and no tracking, validity of funds or anything. and only for 1 'block' a day logging
if this was done every 10 minutes still without all the movement logs thats 144 times bigger
thats 1.7petabytes
(you definetly should be seeing the advantage by now)

and yet blockchain can store all the validity of all transactions throughout a day, every day and every year for 10 years, ensuring no one just magic up new funds with no genesys..  all for less than 250gb,
along with storage saving. the data is more complete with chain of transaction movements during the day, its also auditor and IT security staff saving too. and no special HR department looking into staff pc usage.

so thats the data size and data security, human costs and auditing covered.(all explain more about the security/human involvement below)
next is the point you made about mysql data base with permissions set that only a certain group can maintain the database


this then puts a central authority that needs to be there to approve/revoke abilities.
imagine a company was founded and they employed a CEO while the founders became board members
those people would spend hours discussing who has what powers who can edit the single datafile, who can do this do that.
to prevent the CEO from going rogue the board mmbrs would ned to be watching the CEO's every keystroke to prevent rogue things happening. as they obviously wouldnt want the rogue act to occur and then handle the aftermath.

in a blockchain this is not nessesary. no one can edit the blockchain, because all locations reject versions/edits that dont fit the majority.
so if a employee went rogue. its not like the mysql system would prevent it. instead the damage is done. the damage is noticed(too late) the rogue employee then has their privileges revoked and fired.(AFTER the damage is done)
in blockchain a employee wouldnt even be able to be rogue to even cause the damage first off. instead the system will maintain itself and keep good clean records of what should be there,where the rejected edit would get reported. thus employee fired without causing the damage.
this takes IT guys to maintain a watchlist. auditors to audit the file, HR to audit the permissions list, HR audit its staff,

as an example, bitcoin does this with its ban ip function. if it receives bad data that doesnt follow the rules the data is rejected and the malicious node is banned from future relaying. all without any human needing to tap their finger
this is something mysql database systems do not offer

remember blockchain is not just a database (blk file) blockchain is a technology that secures the data without human decision/power


no
BITCOIN was made to be global
BITCOIN was made to be adversarial
BITCOIN had an incentive mechanism to promote competition
but BLOCKCHAINS dont need to be global, nor adversarial

blockchains solution to the byzantine generals rule is to be mutally in agreement(consensus)
the adversarial part is to reward people for putting more computing power into making a ever increasing harder hash.
take away the incentive people wont increase the hash hardness. but the evr increasing hash hardness is not an essential fature for a private blockchain nor is an incentive

bitcoin is not a good example of blockchains full utility/ability anymore because bitcoin bypasses many critical checks and allows things to be slid in without consensus. not only that but bitcoin is only a small niche usecase of all the possibilities

take airport security(passenger lists/passports/freight/customs/border control). they have to inter-communicate across countries, but dont trust each others countries so want to secure data independantly. not to be adversarial, but to mutually coexist so that the word "trust" does not even come into the conversation

EDIT 2
the UTXO set of bitcoin is just one dataset instance of current balance at one given time. and it is 3.5gb right now. if backing up this dataset of balance every 10minutes would produce terrabytes over time. and still not have the auditing of how the funds changed ownership

it has nothing at all.. nothing . zllch to do with core crippling the blocksize.

if blocks were 0.1mb (10%) then over the same period a traditional database backup of account balance just once a day would be 1.2trrabyte and the blockchain would be 25gb
emphasis on the difference between 25->1200.. emphasis on the 48x saving
emphasis on the 250->12000 emphasis on the 48x saving
again for emphasis 48x saving by doing it on blockchain
emphasis of 48x saving no matter if blocks were 0.1mb or 1mb.
again. that 48x saving is with the traditional database backed up just once a day not showing all movements
if done every 10minutes the saving would be 6912x no matter what the blocksize comparison would be

so dont twist it into a story how cores crippling the blocksize causes the 6912x saving because the 6912x saving would be the same thing no matter the blocksize

I'm only trying to learn why it has to be "a blockchain", if it's private and centralized, and that a database would be more effective and efficient.

Is a centralized blockchain more efficient than an append only database? Because everyone wants to ride the blockchain bandwagon.

thre is no law OFCOURSE to say that people need to convert to blockchains. so thats why i avoided answering your obvious question.
but when databases are append orientated rather than edit orientated, and when databases are distributed and need a QUCIK way to compare versions with each other without the checks themselves taking up all the bandwidth/cpu time. then blockchains have many advantages

in the future there will b hybrid database structures that dont quite follow the full definition of a blockchain. but take aspects of the distributed security and check mechanisms and apply them to traditional database models

people are already doing this with things like sharding and other concepts. where its just not one single blockchain containing everyones data but just a summary chain that each region can use to authenticate another region passes certain tests
EG imagine the international fiat reserve(international monetary fund(IMF)) that has the movements of countries total holdings but not listing every bank holders movments within those countries.

people think blockchains have to continuously go on forever.
what people dont realise is blockchains can b invented for just a 365 day use. take the nations tax account. able to validate which government departments and public grant establishments the funds get moved to. and at the end of the year. its summerised and any excess is moved to a new blockchain. where via another blockchain that just summerises the in-out of each year-end/new year budget

but back to your question.
its like you are saying should all fruit be converted to bananas and all fruit should be a curved at a certain angle, yellow with no black spots. as a weird way to then say that bananas are the devils fruit and you think bananas should not be used/accepted in any way, shape or form
because im getting a real big vibe that you simply hate blockchains.. or just yet to understand them even after a couple years oppertunity to research them

You didn't answer the question.

Do distributed systems' databases have to structured like a "blockchain", or not? Plus do they really need to be "append-only"? But, in the context of the topic, they're centralized.

of course they are not new... dang.. research harder.
the whole point of blockchain is not about the distributed database. its not about the contents and rules of validating each pice of data insid a block/database
 its about while having a distributed database, checking that everyone has the same thing. and done so very simply.
by announcing the latest block hash, which if all other locations have the same hash. means everyone has the exact same data. so simple

block hashes have other uses too. such as if hashes dont match. checking the previous hash, and then the previous hash to the point where they do match. and then coming back one block where it didnt match, to then see whats been changed.
makes it childs play to find out who or whats been changed without having to check every entry inside every block

however if it was just some single file database with a single hash that changes at a periodic time(traditional database). then the only thing a database would have is the latest EDIT and unable to see when things started to change

many companies keep backups of previous instances of a traditional database for that solution. but trying to compare each instance and find out when a change occured is harder and more of a manual task. plus the storage of such is much higher

imagine the data required for every 10minute instance change of the 'bank balance'/UTXOset, a back up was done. yep backing up bitcoins utxoset each time it changes is 3gb extra, which needs storing every 10 minutes. this is far worse if you add up 10 years of it.
3gb x 500,000 instances =15petabytes
yep if bitcoin didnt have a blockchain and just had the utxo set and auditors want a way to check the history and find out when changes happened on certain instances. it would require 15 petabytes and alot of manual checking.. yet blockchain does it all and at a cost after 10 years of under 200gb.. while by default offering all the easy validation, auditing and whatnot built in

franky1, I understand your side of the debate. But do distributed systems' databases have to be structured like a blockchain?

Because "distributed database" systems aren't something new.

your missing a few basic points here.
1. imagine you had blocks 1,2,3,4,5,6,7,8,9,10.. in blockchain you cant just edit an entry in block 5. without it affecting 6,7,8,9,10 and thus other nodes pick up that one version doesnt match the majority
as thats called editing. what you are describing is a complete history destruction and rewrite.. theres a big difference. blockchains dont allow editing

2. for a complete rewrite would require the whole node set to comply. by which they obviously havnt bothered making a true blockchain with all the checks to prevent it by default

3. if you dont understand the whole concept of the chained hashes, the distributed verification, checkpoints and such, then its time you looked into blockchains

4. i see the purpose of blockchains in different industries done so, so that temp employee at random bank branch cant just add a few zeros to the end of his balance and then run off to a tropical island,, rich.

5. an im pretty sure if a company wants a secure data system then they would have systems inplace to make sure it actually a blockchain and not some backdoor filled crap that can erase data from all systems at the same time.

even things like handing a checkpoint hash periodically to some attorney who can, if any data integrity quieries arise, can check to see if its true or not. which is backed up by some contract that fines the CEO personally if such event arises

"Editing" as in "writing data to" (eg. registering and timestamping the image hash, patent, prescription as per the use cases mentioned above).


Speaking of editable:



Like I said:


Or put differently: If the CEO of a company (or director of a government organization) tells their IT to roll back the private blockchain, that's exactly what they'll do: Collaborate to change supposedly immutable data.

editable?.. you must by now understand blockchains are append only, i mean come on its not your first day on the scene
first understand distributed blockchains.

even in a system where all the nodes are run on a companies intranet(internal network) its still not possible for a single entity to edit the block/data. it would require all the employee's to collaberate.
hense the need for distribution of the data. and the need to then secure the distribution by having a simple verification procedure which blockchains offer by default(hashes)
yes in a central stored single database on a central server can b edited easily.
yes in a distributed single database without any security checks can be edited easily.
but to change data on all locations simultaneously whereby each location can check its peers for such attempts is something that single individual cant do.. plus it dosnt require IT guys, security teams, auditors to keep a check on it. as a blockchain is self checking/auditing

alot of people think centralised is not able to be distributed and decentralised cant be controlled by a certain group
hense if you get off the buzzword of 'centralised/decentralised' and think about blockchain distribution instead, you will soon learn what blockchains are really about, whether it be centralised or decentralised

That was not a general question but rather referring to the concrete examples mentioned above:




I'm questioning the applicability of private permissioned blockchains for these specific use cases.

If we get back to the flow chart again, the point is that e.g. different companies work together (write) that do not trust each other.


You have a consortium consensus in permissioned blockchains, where a Tx is sent to pre-selected nodes and the consensus validates the Tx.


That's true.


Let's just agree that the use cases of a permissioned blockchain are really very limited and there are enough alternatives available.  


Hyperledger would be the Project and Hyperledger Fabric for example one of the blockchain frameworks.


Is it a distributed chain of data linked using cryptography?  

This use case (ie. timestamped verification) has been widely discussed for years and to some extend even been put to practice -- on public permissionless blockchains.

What about private permissioned blockchains though?

How is a private permissioned blockchain run by private company nodes, editable only by company employees more trustworthy than the same set up using classical databases and access control? Same for private permissioned blockchains run by government entities on government nodes, editable only by government employees. Where's the benefit in this case?

Then the question, "Does the database have to be a blockchain?", or do those companies want a blockchain for the buzzword?


Did I say it? No. A "blockchain" implemented correctly is censorship-resistant, trustless, and decentralized.


I'm very unprepared, yes. Newbies, listen to franky1, he is the most prepared.

But, does it have to be a blockchain?

alot of people think that blockchains wont work for companies because people believe a blockchain has to not only be a single database. but also has to include all the data.

imagine a blockchain where each block doesnt contain whole image files. but just the file hash of an image file, where by the image is stored separately.

people can then validate the images are authentic/not been edited because they can see if the hashes match the ones in the blockchain.
imagine its use for patent registration. the patent office doesnt need to store all the patent wording/images in the database, just the file hash of the submission. and then because the file hash is in a blockchain. its then timestamped and logged. anyone trying to edit a patent file ends up with a different file hash. which when comparing it to the blockchain. wont match and people can see that its a false/edited document

this then allows everyone to have a distributed database, without the worry of having to store millions of peoples files aswell, especially if they dont personally need to know about others files content

same goes for things like pharmacy prescriptions. a blockchain does not need to store the patients name, address, social security, doctors details, and medication+amounts. it just needs to store a hash of the data. which the pharmacy can then check actually exists in a blockchain. those know its a real prescription and not one created by a druggy using photoshop

1. companies do need append only databases
take your car log book. you think it only lists you as the owner.. ever.. no. the car registration authority appends each new owner when the ownership changes. just because the latest append shows as you, doesnt mean that no other entries exist.
same with land registry and many other things.
also going into financials. like banks. ever thought about bank statements and how your only comparison of a blockchain being bitcoin is the same thing. where a wallet logs all payment history of an address/account
ever thought that a bank balance is just a calculation done after adding/subtracting the total append only payments in and out of an account... think about it how else do you think they get to the 'current balance' if they are not taking previous ins and out into account

so "why would a company want append only." .. simple. to keep a log to check everything audits correctly
so "how to prove 'private-centralized-blockchain-database' is immutable".. simple. distribute it and check the hashes.. you really need to understand how blockchain security works.. (hint: do research)

2. by you saying blockchains cant scale, blockchains dont work, no one wants blockchains.. yes you say those things. you have been very very visible in your hatred of the technology. your only interest is the price of the asset and the desire to centralise it under a certain groups control. i would feel sorry for anyone that tried learning things from you not only for your motives, but also for your own lack of independant research

3. you are so unprepared to understand blockchains that you cant even recognise that blockchains exist even when one is being described to you. here il describe it again:
"imagine a distributed database, which by its own design backs up each other (if one goes down others are available) where by design individuals cant tamper with it, it doesnt need IT guys, security, auditors either."
also if you dont think these databases exist. then you are saying bitcoin is a figment of peoples imagination!?!
i think its really time you start doing your research. not just about what blockchains are/do/how they work. but also about general life stuff like companies that can use such, are using such and will use such.

have a nice day, but try not to spend it just replying. spend some time doing research first. it will help you

i do have to laugh that you honestly thought such databases didnt exist.. that made my day

Is Hyperledger a blockchain? Is then there a need for a blockchain, if it's centralized? Took take it further, is Bitcointalk forum then a blockchain? I'm confused.

That's the thing though, so far it seems like permissioned blockchains don't provide consensus, immutability and double-spending protection either. They provide the illusion of such.

Say a company runs their own instance of the Ethereum blockchain. What prevents them from rolling back unwanted parts of history?

After the public Ethereum blockchain rolled back / hard forked due to the whole DAO debacle the aftermath at least resulted in some purists running the original ETC chain. A lot of people needed to decide in favour of ETH over ETC for ETH to prevail -- when choosing which chain to mine, which codebase to work on, which currency to buy.

In a permissioned setting rewriting history becomes a mere executive decision.

Don't get me wrong. I'm very curious to see where decentralized consensus can lead us beyond currencies. But I think the use cases in non-public and permissioned settings are extremely limited.

I totally get your point and I do agree with you that git and blockchains are similar in many ways.
Yet they differ for a number of reasons.
For example: git doesn't provide consensus, history can be re-written, it doesn't prevent double spending ...
Just to name a few.

We could also compare blockchains with other p2p databases, like distributed hash tables.
Every system has its benefits and drawbacks.

Why is everyone here always thinking that permissioned blockchains have no right to exist?
Hyperledger would be a good example here.


To persist/save state 

Let's stick to the hyperledger example.
Here, immutability is ensured by the so-called ordering service.

An append-only database. Why would companies want that, and if they truly did, how can they assure that their "private-centralized-blockchain-database" will be immutable, and secure from tampering?



Troll harder.


I'm the stupid one. You are the person that newbies should listen to. I hope they do, and learn, the hard way.


How can tampering then be assured impossible? What "proof" can they show its users that it's not "tamperable".

Is there such a "database"?

when it comes to wanting to edit or change a certain piece of data by multiple users, then yes alternatives exist. but when it comes to needing to retain a secure copy of data that cannot be edited then thats where blockchains strengths are.

like bank transfers and retaining bank statements for years
parcel/goods delivery which can trace back to source (food distribution industry love this)
car manufacturing/maintenance logs
hospital records
education/student grades
land ownership authorities
property management/repair
intellectual property authorities
insurance
.. the list goes on

im not saying everyone has to convert to blockchain, as thats stupid. but many in this topic are acting as if blockchains are useless and impractical for any use.