News for hackers. Stuff that matters.
Re:Blockchain, the Solution for Almost Nothing by andreskytt on Saturday August 22, 2020 @01:55AM (#24241978)
Number mobility between Indian telecoms runs on private blockchain. AFAIK it is live and actively used. BC solves a problem there because operating a database for that scale would be very complicated. So every telco has their own copy that get synced via a consensus mechanism and signed transactions
Re:Blockchain, the Solution for Almost Nothing by Lazare on Saturday August 22, 2020 @02:21AM (#24242101)
> private blockchain

So a database?

> AFAIK it is live and actively used.

Most things people seem to think are running on blockchain tech turn out not to be, so I'd be interested in any evidence here. Nonetheless.

> operating a database for that scale would be very complicated

They're already operating a database, so at most this is an argument that a traditional database tech wouldn't work, except...

...this is only a few billion rows, at most, right? With minimal data? It would totally, totally work. We're at a scale where a copy of "MySQL for dummies" is going to explain everything you need to know here. A moderately large number of people in this thread probably work for companies that have traditional databases that work at a larger scale or with higher demands than this.

Re:Blockchain, the Solution for Almost Nothing by davidgerard on Saturday August 22, 2020 @05:31AM (#24242929)
> Most things people seem to think are running on blockchain tech turn out not to be, so I'd be interested in any evidence here. Nonetheless.

There's a few such systems in production - in practice, they tend to use a private Ethereum or Hyperledger as a back-end data store, and 100% of the use case for using a blockchain for your private back-end data store is being able to say "blockchain!" in the press release.

Re:Blockchain, the Solution for Almost Nothing by an_opabinia on Saturday August 22, 2020 @03:17AM (#24242360)
I am not super familiar with the particular example.

But if you consider it more like an easy to cluster database with free, tamper-proof backups, it makes a lot more sense.

Re:Blockchain, the Solution for Almost Nothing by redis_mlc on Saturday August 22, 2020 @04:43AM (#24242727)
I've worked with MySQL and Postgres databases that invariably include one or two tables with over a billion rows per table. Typically the rows record user fingerprint info of some kind, either UA or location.

There is no problem with performance until schema changes are made, requiring a careful migration strategy plan (this is why DBAs get the big bucks.)

Source: DBA.

Re:Blockchain, the Solution for Almost Nothing by csande17 on Saturday August 22, 2020 @02:13AM (#24242060)
> BC solves a problem there because operating a database for that scale would be very complicated.

Would it, though? Google tells me that India has a population of around 1.4 billion people. Let's assume each person has one phone number, and each phone number requires one kilobyte of number-mobility-related data. Then the database would need to store 1.4 terabytes of information.

1.4 terabytes doesn't seem like an impossible amount of data to store. That fits on a consumer hard drive! Dell will sell you a single server with that much memory! Request volume doesn't seem like a huge issue either, since porting a phone number is something people do extremely rarely.

Re:Blockchain, the Solution for Almost Nothing by mongol on Saturday August 22, 2020 @03:04AM (#24242305)
I don't think it is the size that is the problem. I imagine it is like a routing table, keeping track which number is managed by which operator, and if India has number portability, there needs to be a transactional method to update it that is shared by all operators. I could imagine this being done with a blockchain.

Or for that matter, recording of call details for billing between operators?

Re:Blockchain, the Solution for Almost Nothing by csande17 on Saturday August 22, 2020 @03:21AM (#24242375)
"Keep track of which number is managed by which operator" reduces to "store a small amount of information for each phone number".

In a traditional database setup, the "transactional method to update it" would be sending a request to a central server to update the database. The central server can be run by whatever regulatory body or industry association is responsible for number portability -- ideally, this would be the group with legal authority to overrule whatever the computer system says anyway. It wouldn't need to handle a large number of transactions per second because people don't change phone providers very often.

You could imagine doing it with a blockchain instead, but you could also imagine doing it using a Yahoo Groups email list -- that doesn't make it a necessary or good solution.

Re:Blockchain, the Solution for Almost Nothing by splix on Saturday August 22, 2020 @12:04PM (#24245003)
What if other operators don't want to introduce an additional governing entity with a super power of taking their customers? Like if they don't want additional corruption, easy to misuse power, etc?

I mean, I can imagine government wants this, so if such ideas comes from government you're going to have such regulatory body. But otherwise, if it comes from operators itself, they don't want to introduce such entity, and they don't want to award one of them as an ultimate boss to rule all of the numbers. Blockchain, at least, gives them a chance to continue to be equal.

Re:Blockchain, the Solution for Almost Nothing by DamnableNook on Saturday August 22, 2020 @06:56PM (#24247942)
Blockchain backers seem to take it as a given that centralization is bad. But when pressed on why, the best answers I have seen are “because centralization is bad”. Well, I guess that’s not 100% true; the first answer is that we shouldn’t trust authority. Why? Because authority is bad. Why? Because authority is bad.

But as this article articulates so well, that’s a red herring. Like, you can come up with all sorts of wild scenarios where central authority can be bad. But ultimately, these are scenarios you’re unlikely to encounter in the real world, and if you do, there are much simpler solutions.

Blockchain technology always reminded me of that XKCD about cryptographer expectations vs. reality (https://xkcd.com/538/): it’s solving a problem that only really exists in fantasy for 99% of the population.

Re:Blockchain, the Solution for Almost Nothing by ric2b on Saturday August 22, 2020 @08:22AM (#24243633)
Sounds good at first, but now your calls stop working if the regulator's system goes down for maintenance. You can certainly make it work, but I don't see what the major benefit of using a central database is here.

> It wouldn't need to handle a large number of transactions per second because people don't change phone providers very often.

That makes it fit blockchain quite well, actually.

Re:Blockchain, the Solution for Almost Nothing by zucker42 on Saturday August 22, 2020 @03:19AM (#24242367)
I imagine the issue is maintaining consensus between the different phone companies. Without blockchain you'd have to have a canonical database run by a single entity, whereas you can use blockchain as the consensus mechanism among competing companies.
Re:Blockchain, the Solution for Almost Nothing by csande17 on Saturday August 22, 2020 @03:24AM (#24242389)
It's a shame there's no central authority responsible for regulating telecom in India, or this would be a really easy problem to solve.
Re:Blockchain, the Solution for Almost Nothing by zucker42 on Saturday August 22, 2020 @03:41AM (#24242450)
I mean sure you could have the government manage the central database but maybe the Indian government is not equip to run such a database. Plus you can avoid some of the problems with high costs or stagnation that can happen with centrally run systems. I don't think there's any particular reason to reject blockchain if it works in this case.
Re:Blockchain, the Solution for Almost Nothing by petra on Saturday August 22, 2020 @03:41AM (#24242449)
Start non-profit for that purpose that will manage the database.

Ther database will hold only entries signed by telco companies. Some rules will apply on what happens when a number changes.

Everyone gets access to the raw database, and to query the database.

In case of mismatches there will be an agreed arbitration process, with agreed upon fines to the party no being responsible. This or some other contract.

This sort of solves the problem.

The key win in the blockchain - is the decentralization. But we know, in the real world how to arrange cooperation using centralization. it works. It's done every day. So this case is mostly useful when you want to avoid regulation.

Another win of the blockchain - a lot of work has been done verifying that it is secure. It probably is the mose secure system for those sort of problems. And that's a good reason to use it in many places.

Re:Blockchain, the Solution for Almost Nothing by csande17 on Saturday August 22, 2020 @04:08AM (#24242558)
I'm not sure I agree that blockchain is the most secure solution for common problems.

Let's say you get a computer virus that steals your banking credentials and transfers your money to someone else's account. In the traditional banking system, it's often possible to reset passwords and reverse transactions, especially if you catch the bad behavior quickly. With Bitcoin, your money is gone forever and it is mathematically impossible for you to ever get it back.

Re:Blockchain, the Solution for Almost Nothing by ric2b on Saturday August 22, 2020 @08:26AM (#24243652)
That's true for Bitcoin because it's almost impossible to get most miners and nodes to agree to reversing such a mistake, it's completely against the spirit of the project.

But in the example above that could still be easily fixed, you'd just need the companies involved to agree to reverse the mistake, as they are the miners of the chain.

Re:Blockchain, the Solution for Almost Nothing by joosters on Saturday August 22, 2020 @02:07AM (#24242035)
You know that a private, or permissioned, blockchain is pointless, right? If some entity controls the blockchain, you might as well just use a standard database.
Re:Blockchain, the Solution for Almost Nothing by ric2b on Saturday August 22, 2020 @08:29AM (#24243669)
It's not pointless, it's just yet another way "noSQL" database. It certainly doesn't give you the same benefits of a public blockchain.

But if the code is already written, meets the project requirements and is easy to deploy, why not?

Re:Blockchain, the Solution for Almost Nothing by kls on Saturday August 22, 2020 @09:52AM (#24244141)
Because it is a really inefficient way to store what amounts to an immutable ledger in a traditional database. While I don't agree with a lot of the consensus that blockchain is completely useless. I pretty much agree that it is useless if it is owned by a private entity. That was one of the main points of the blockchain, how can we create a public datastore that is not owned by anyone and it tamperproof. I see public blockchain as valuable to freedoms in a libertarian sense but see no value in companies adopting the core blockchain techs. Tooling around public blockchains sure or API's available on things like D'APPS make sense, but private ledgers make absolutely no sense to me. Just use a traditional database and implement all the data storage as a ledger pattern and you get the same effect with out all the energy consumption.
Re:Blockchain, the Solution for Almost Nothing by splix on Saturday August 22, 2020 @12:13PM (#24245074)
It isn't owned by a single private company, most usually. But multiple companies with equal rights to update. That's what you cannot do with traditional db.

Also I don't think there is any significant energy consumed. There is no mining or anything like that. You have to pay for extra network traffic and more disks, simply because you're going to have more copies of data, most likely, but it's not so large amount of energy.

Re:Blockchain, the Solution for Almost Nothing by joosters on Sunday August 23, 2020 @01:59AM (#24249904)
An API doesn’t care what company the client belongs to, so it’s pretty easy to have a DB that can be updated by more than one company. There’s no technical reason to prefer a blockchain, it’s just marketing.
Re:Blockchain, the Solution for Almost Nothing by splix on Monday August 24, 2020 @10:51AM (#24261232)
It's pretty easy to make it like you said. But it also pretty easy to make it in a way when some companies would have a preference.

Second would be more profitable for the owners of that db, so there is a motivation to do it in that way. In many of the countries outside of US/EU it's a popular practice.

Re:Blockchain, the Solution for Almost Nothing by ric2b on Saturday August 22, 2020 @11:56AM (#24244946)
AFAIK private blockchains usually don't rely on proof of work so there's no large energy usage.

I agree that if they're private they're equivalent to a customized database system, private blockchains aren't interesting but it doesn't make them useless, it's simply another way to solve problems that could already be solved before.

Re:Blockchain, the Solution for Almost Nothing by smsm42 on Saturday August 22, 2020 @02:10AM (#24242048)
This looks like just a replicated append-only database. One can call it "blockchain" but it's not exactly anything new.
Re:Blockchain, the Solution for Almost Nothing by sohamsankaran on Saturday August 22, 2020 @02:05AM (#24242028)
Reference? AFAIK this was proposed by Wipro in a white paper but was never implemented.
Re:Blockchain, the Solution for Almost Nothing by Beldin on Saturday August 22, 2020 @06:00AM (#24243044)
Number portability in my country is handled without blockchain. Granted, India's population is about 100 times larger. But I don't see a reason why a scale-up with a factor of roughly 100 would be prohibitively problematic.
Re:Blockchain, the Solution for Almost Nothing by splix on Saturday August 22, 2020 @12:23PM (#24245145)
Maybe it's not about transaction scale. And I don't really see how blockchain helps with scale, imho it's opposite.

With this particular case, it maybe be solution to fight corruption (just a guess). Centralized trusted control of all phone number of a country is a huge power and a magnet for different types of corruption. So maybe the difference is that in your country the magnitude of corruption is lower than in India, or easier to handle because of size.

Re:Blockchain, the Solution for Almost Nothing by Beldin on Wednesday August 26, 2020 @03:38PM (#24286380)
> So maybe the difference is that in your country the magnitude of corruption is lower than in India,

I don't know about corruption levels in India. I am under the impression that the level of corruption in my country is rather tiny, so this could very well be the explanation.

Re:Blockchain, the Solution for Almost Nothing by pgo on Saturday August 22, 2020 @02:06AM (#24242032)
Source for this ? I can only see mentions of plans implementing this, not of it being actually deployed