3 ms·
> Or, if you’re running a large logistics company with one or more supply chains made up of many different vendors, and need to identify, track, trace, or sourc
by dnautics 8y ago
> Or, if you’re running a large logistics company with one or more supply chains made up of many different vendors, and need to identify, track, trace, or source the items in the supply chain, then blockchain may be the solution for you as well.
Can someone explain to me why. You need a block chain for this? Why not just an evented database with a signature field and periodic hashes?
- ssalka 8y agoHow would you prevent a rogue developer from updating/deleting events? Even if you could tell from the hashes that something changed, you would lose whatever data the dev changed. Backups can help mitigate this, but not until after the fact.
- larrysalibra 8y agoThe same way American Express, Visa, the Automated Clearing House and your bank prevent rogue developers from updating/deleting your money: With internal security processes, threat of termination and ultimately the legal system.
- syrgian 8y agoHave database replicas that only listen to INSERT statements (easy to implement in My/Maria SQL with BEFORE triggers), and design the software that works with the DB as append-only.
- icebraining 8y agoRight. Or any other append-only system, really. I mean, one wouldn't ask "how do you prevent someone from deleting an email they sent you earlier?"
- the_gastropod 8y agoNot necessarily. Think of git. Git is not a blockchain. If a rogue developer deletes code and/or rewrites history and forces a push, I will know and can undo that rather easily.
- akerl_ 8y agoAmusingly, that's because git is a merkle tree, which is one of the core building blocks (ha!) of a "blockchain" as it's used by "blockchain" enthusiasts.
- larrysalibra 8y agoA blockchain doesn't handle the supply chain problem very well because supply chains + blockchains suffer from the oracle problem. Blockchains don't know anything about information outside of their chain and need a trusted 3rd party - an oracle - to tell them about it. Imagine you want to track eggs through the supply chain from farm to table. Ultimately the trust of your system is rooted in the farmer and various other organizations that are updating the state of these eggs in the chain. If they make up fake data, the system isn't trustworthy. Typically the solution for this is to only let trusted players into the system and have some sort of organization - perhaps the government - that vets users of the system and penalizes those that enter fake data. In this case, the trust of your blockchain-egg-supply-chain-tracking system is rooted in the organization accrediting and inspecting users of the system. If that's the case, you're better off trusting that organization to run a traditional database. You shouldn't use a blockchain.
- vidanay 8y agoThis is akin to the ISO-9000/9001 problem where you can produce a very shitty product, but it's an exceptionally well documented shitty product if you ignore the corrective action side of the equation.
- rstuart4133 8y ago> Can someone explain to me why. You need a block chain for this? We have been using analogues of the system described for literally centuries. A titles office is very like a block chain. Rather than currency being transferred between people, it's titles to land or buildings. Individuals do transactions by creating a transfer and giving it to the titles office to be blessed. The transfer is worth nothing until the titles office puts it's stamp on it. The titles office function is to create an immutable, consistent chain of ownership. The formal name for the software equivalent is a "permissioned block chain". There is nothing particularly special about it - if you gave the problem to a bunch of 4th Uni Students, this is what you would expect to come back. The term "block chain" looks to be a marketing grab to me. I don't think it really applies - the only thing they have in common with crypto currencies is they record a chain of ownership. Crypto currencies have blocks so they can spread the high cost of proof or work over many transactions but there is no expensive proof of work in permissioned block chains, so they don't need blocks either. All that aside, given we have been using these manual systems for centuries I don't think there is much doubt automating them with electronic equivalents would be result in something that is both faster and less costly. But whether it's a good idea for logistics chains is debatable. One thing the titles office and crypto currencies necessarily have in common is there there isn't a lot of "new stuff" entering the system - they are tracking existing stuff (the titles or currency), which they can easily verify by just checking the person authorising the transaction is the person recorded as the current owner by the chain. When new stuff does enter the system it is very tightly controlled. Crypto currency does it via mining. Titles are usually created by acts of parliament releasing land. In these logistics systems new stuff is continually entering the system, and some separate external system has to vouch it is in fact a tray of A grade fruit or whatever. But I sure whoever is fogging their existing permissioned block chain implementation is telling every customer with a vaguely similar system it is a near perfect fit.