Distributed database project

[_green_lion_]

Hello,

I have a project that I believe will attract attention. I've been in software business for a long time. We are in a time when the IT industry is heading for the cloud, and my project that has puzzled my mind is just right for this job. Let me briefly talk about my project. You know, cloud sql is a very advantageous system. Companies such as Google, Amazon, and Microsoft are very good at this. Many companies are trying to exist. Bitcoin and decentralized systems excite me. I want to make the cloud SQL system a distributed application and I think I have the software knowledge to do this.

The system will work like this.
1 - There will be customers who are members of the system and want to use SQL with reasonable pricing.
2 - There will be data miners who want to share their computer's processing power and bandwidth.
3 - As Gateway, our software will provide communication. It will redirect the SQL system to idle miners like clusters.

Our solutions are ready for some problems that may occur in the system. For example, what to do if miners shut down their computers or are unavailable In order to avoid such a situation, the data of each miner will be available to 4 miners at the same time. As soon as the miner cannot be reached for a while, the dbs in his possession will be matched with other miners and the unreachable miner will not be able to receive new customers until they gain trust again. It will work with redundancy like a Raid structure. Also, miners will not be able to read their db. Databases will be encrypted with high encryption.

In other words, people who have old and unused computers may earn up to $ 200-1000 monthly passive income. There may be much more income depending on the customer's preference.

I also intend to support my project with a coin or token. I want the money I will pay to the miners and the money I will get from the customer with this coin.

I intend to cut a certain percentage of the earnings of those who mine as a source of income in the system. All other remaining money will be transferred directly to the miner as coins.

In your opinion, what kind of path should I follow to implement this project? Or do you think my project arouses excitement? I will be waiting for your valuable comments.

[1715443659]

1715443659

[1715443659]

1715443659

[1715443659]

1715443659

[joniboini]

As far as I can remember, there are dozens of projects out there trying to solve the same problems as yours, albeit with different focuses. I think issuing a new coin is not necessary since you can simply ask for a payment from the customers and use it to rewards the data hosters.

Not everything requires a new token, you can simply use Bitcoin or other coins if it is only required for payment.

[NeuroticFish]

I agree with joniboini here. You don't need your own coin/token for this.

You can make some sort of "shop" where some will "order" work, pay for it and you can take your cut and which also acts like a mining pool and based on the work done, the accounts of the workers will accumulate money. And that money can be in any form, from Bitcoin to major altcoins or even fiat.

The problem you may have to think about would be what happens when you have "miners" and no orders. Since the miners may find some other work to do if you don't have anything to offer and they may be gone for good.

[_green_lion_]

I agree with joniboini here. You don't need your own coin/token for this.

You can make some sort of "shop" where some will "order" work, pay for it and you can take your cut and which also acts like a mining pool and based on the work done, the accounts of the workers will accumulate money. And that money can be in any form, from Bitcoin to major altcoins or even fiat.

The problem you may have to think about would be what happens when you have "miners" and no orders. Since the miners may find some other work to do if you don't have anything to offer and they may be gone for good.

Thanks for your valuable comments. Here comes the logic of generating coins. Data miners can collect some of the coins we produce when there is no demand. So it's kind of like an airdrop process. I think we can implement a policy to prevent miners from exiting the system with this transaction. This is exactly why we want to produce coins. Would it be a logical action in your opinion?

[NeuroticFish]

Thanks for your valuable comments. Here comes the logic of generating coins. Data miners can collect some of the coins we produce when there is no demand. So it's kind of like an airdrop process. I think we can implement a policy to prevent miners from exiting the system with this transaction. This is exactly why we want to produce coins. Would it be a logical action in your opinion?

Having your own coin is a tricky direction. You have to advertise it. You have to make it valuable. You have to pay exchanges to trade it. It can easily become the failure point of your project.
I would rather divert the miners do different lucrative tasks that can still produce money (for example actual mining) instead of coming with my own coin. But it's really up to you, it's your project.

[_green_lion_]

As far as I can remember, there are dozens of projects out there trying to solve the same problems as yours, albeit with different focuses. I think issuing a new coin is not necessary since you can simply ask for a payment from the customers and use it to rewards the data hosters.

Not everything requires a new token, you can simply use Bitcoin or other coins if it is only required for payment.

If it won't keep you busy, can you share such familiar projects for me to review? Maybe it can be an idea for us or we can contribute to those projects.

[_green_lion_]

In other words, people who have old and unused computers may earn up to $ 200-1000 monthly passive income. There may be much more income depending on the customer's preference.

Since you mention old/unused computer, what will happen if there's bottleneck which limiting speed of accessing the encrypted database?

We also thought about that part. As a matter of fact, it is certain that it will be more performance than SQL servers deployed on virtual machines on google or aws. Nevertheless, given the performance request of the customer side, we made the following logic.

When each miner is added to the system, the hardware specifications and performance graph will be monitored continuously. Miners with the best device and network width will be at the top of the rankings and will be the first to catch the customer.

Since each database is in 4-5 miners like raid logic, whichever miner is in the best state at that time, the query will be answered from that miner. The miner who answers will also earn the fee for the query.

Just like the rule that establishes the best system in mining wins the best money, the most powerful computer and bandwidth will respond to more requests and earn more.

[LoyceV]

You had my attention right until this part:

In other words, people who have old and unused computers may earn up to $ 200-1000 monthly passive income. There may be much more income depending on the customer's preference.

This doesn't seem to be very realistic. An old computer is worth less than $200, if it can earn more than that, anyone can build a farm. Or even easier: an old computer can be compared to a $10 VPS, and if that can earn 20-100 times more, it's very easy to turn on thousands of them and earn a million bucks per month.

Who's going to pay $200 or more per month for a task that can be performed by a single old computer? Or even worse: they'll have to pay that amount 4 times, because there are 4 copies all wanting to earn the same monthly income:

the data of each miner will be available to 4 miners at the same time.

I can think of some uses for a distributed storage system, such as storing settings and small amounts of other data for a non-persistent operating system. But it has to be cheap to use, and you should expect earnings per miner to be limited by market forces.

[_green_lion_]

You had my attention right until this part:

In other words, people who have old and unused computers may earn up to $ 200-1000 monthly passive income. There may be much more income depending on the customer's preference.

This doesn't seem to be very realistic. An old computer is worth less than $200, if it can earn more than that, anyone can build a farm. Or even easier: an old computer can be compared to a $10 VPS, and if that can earn 20-100 times more, it's very easy to turn on thousands of them and earn a million bucks per month.

Who's going to pay $200 or more per month for a task that can be performed by a single old computer? Or even worse: they'll have to pay that amount 4 times, because there are 4 copies all wanting to earn the same monthly income:

the data of each miner will be available to 4 miners at the same time.

I can think of some uses for a distributed storage system, such as storing settings and small amounts of other data for a non-persistent operating system. But it has to be cheap to use, and you should expect earnings per miner to be limited by market forces.

Thank you for your comments.

Even if what I think is not fully formed, it seems to be largely ready. Let me try to explain them a little bit.

First of all, I made these price predictions by looking at the prices of other cloud accounts. For example, when you buy an average cloud database system, you pay both storage and traffic fees. These servers are also elected. For example, there are differences in the form of vcpu and hacpu / vcore and hacore. So an average database service costs between $ 20 and $ 40 per month. And if you want hardware cpu and ram, these prices go even higher. Like 100 to 200 dollars.

In the system I mentioned, all user computers are devices with hardware CPU and RAM. Even if its features are not as good as a server, I still think that they will be enough to keep an average of 10 databases alive (some companies I support have 5th generation intel processors and stable databases with 8 GB RAM) The person who opens this type of computer to mining for free, as much as amazon or google Of course, he will not expect to earn big money. The db size stored in the system does not matter much. Companies usually charge the traffic. Since the miner software will constantly monitor the performance of the computer and the network, requests will be directed to the best system side each time, and the best system will earn the highest price.

It is necessary to be cheaper for cloud prices. It is necessary to charge half or less than other central server services. The higher the number of customers, the easier it is for miners to find customers and earn profits.

That's how I think. I really need your advice and criticism. Please feel free to write down any questions you would like to ask whether this information satisfies you.

[eckmar]

Not possible. I think you are not a software engineer yourself, so maybe you should talk to one. If you take a look at the big companies you mentioned and their flagship super scalable database solutions (DynamoDB, Firebase, Cosmos DB) they are all nosql (document databases) and not sql database. Thats because of the simple reason, that relational databases are harder to scale across multiple instances, which is exactly what you plan to do.
Second and most notable thing, performance. One of the reasons people like sql databases is strong consistency for writes. That means that if you send a write request and right after that a read request for same data you would instantly get what you wrote. Beforementioned solutions by big companies usually sacrifice that for performance. In short, for each request you would need to propagate data across all the nodes responsible for hosting the users data, and you'd need to have consensus protocol. Thats all slow. Expected response time for query is in milliseconds not seconds.

I hope you find a solution for this, if you do be sure to message me and I'll be the first one to use your product. I'm also sure you'll get a big fat paycheck by one of the big companies for that solution.
Good luck.

[_green_lion_]

Not possible. I think you are not a software engineer yourself, so maybe you should talk to one. If you take a look at the big companies you mentioned and their flagship super scalable database solutions (DynamoDB, Firebase, Cosmos DB) they are all nosql (document databases) and not sql database. Thats because of the simple reason, that relational databases are harder to scale across multiple instances, which is exactly what you plan to do.
Second and most notable thing, performance. One of the reasons people like sql databases is strong consistency for writes. That means that if you send a write request and right after that a read request for same data you would instantly get what you wrote. Beforementioned solutions by big companies usually sacrifice that for performance. In short, for each request you would need to propagate data across all the nodes responsible for hosting the users data, and you'd need to have consensus protocol. Thats all slow. Expected response time for query is in milliseconds not seconds.

I hope you find a solution for this, if you do be sure to message me and I'll be the first one to use your product. I'm also sure you'll get a big fat paycheck by one of the big companies for that solution.
Good luck.

In addition to the nosql solutions you mentioned, there are also SQL solutions directly. For example

Amazon RDS
https://aws.amazon.com/tr/rds/
Azure SQL
https://azure.microsoft.com/en-us/services/azure-sql/
Google Cloud SQL
https://cloud.google.com/sql

I know very well the difficulty of the subject I am talking about. You are talking about the consistency of the select query following Insert/update from a different peer. As I said at the beginning of the topic, I believe we can do it and I wonder if it is worth the effort.

[eckmar]

Not possible. I think you are not a software engineer yourself, so maybe you should talk to one. If you take a look at the big companies you mentioned and their flagship super scalable database solutions (DynamoDB, Firebase, Cosmos DB) they are all nosql (document databases) and not sql database. Thats because of the simple reason, that relational databases are harder to scale across multiple instances, which is exactly what you plan to do.
Second and most notable thing, performance. One of the reasons people like sql databases is strong consistency for writes. That means that if you send a write request and right after that a read request for same data you would instantly get what you wrote. Beforementioned solutions by big companies usually sacrifice that for performance. In short, for each request you would need to propagate data across all the nodes responsible for hosting the users data, and you'd need to have consensus protocol. Thats all slow. Expected response time for query is in milliseconds not seconds.

I hope you find a solution for this, if you do be sure to message me and I'll be the first one to use your product. I'm also sure you'll get a big fat paycheck by one of the big companies for that solution.
Good luck.

In addition to the nosql solutions you mentioned, there are also SQL solutions directly. For example

Amazon RDS
https://aws.amazon.com/tr/rds/
Azure SQL
https://azure.microsoft.com/en-us/services/azure-sql/
Google Cloud SQL
https://cloud.google.com/sql

I know very well the difficulty of the subject I am talking about. You are talking about the consistency of the select query following Insert/update from a different peer. As I said at the beginning of the topic, I believe we can do it and I wonder if it is worth the effort.

I'm well aware of those, but that services are not marketed as extremely scalable solutions. Yeah its worth the effort if you can make that. If you crack this and make a distributed sql storage that responds in milliseconds, you'll be a millionaire easily. No one has been able to do that so far. Thats the reason big guys invented nosql, so it can scale easily across many instances.

Consistency is just one of the issues, your sql db needs to be acid compliant otherwise, there is no point in using it. You can’t just reuse mysql/postgres for this, you need to make your own db.