How can you create a database that works across platforms?

Understand what the database needs to deliver. Build tables or a structure that holds the necessary data. Plan for future changes. Follow industry standards like 3nf when designing your tables. Once the database is designed, build a good set of REST apis that deliver the basic requirements to the user. REST is cross-platform, and really easy for developers to consume.

For creating cross platform database typically involves DBMS (Database Management Systems) like PostgreSQL,MySQL,Microsoft SQL Server further more it is noted that you can use variety of cross platform development frameworks for connecting DB like Hibernate for Java or Entity Framework for C#. So in my opinion this is still more feasible than ever

Does "AI" know that "nobody has been fired for choosing ... as database product"? What is a database and what a database system?

IDMS!!! After almost 60 years and I shudder to think how much money was spent but there still has not been a database management system that competes with IDMS for mainframe, web, cloud, cell phones, etc.

Why is pay for databases not included? There are benefits to paying for a database and the service they provide. Oracle, SQL sever, come to mind.

Every database is accessed via APIs (Application Programming Interfaces) or SDK (Software Development Kit), that can be direct (i.e. the ones provided by the Database vendor), or indirect (custom services, created by you or other developers). In the first case, every DB vendor takes care of making the API/SKD available for all the major platforms, or in the worse case scenario, they make an HTTPS interface to interact to. In any case, you can consider all the major DB engine to be cross-platform in terms of usage. There are exceptions, of course, but they belong only to the realm of small and fast experimentation, definitely not for any serious deployment.

In other words, an I/O module. These are a very bad idea for database applications. It's like pouring Lake Superior into a funnel. Having every program that accesses the database go through an I/O module will destroy performance and forget making SLA's.

Cross-platform is fine but with Microsoft's support for linux these days, all a good database engine need to support is linux. And if it can run in a container, then popular platforms such as Kubernetes can make running in production a breeze and developers can play with it locally too. Different linux distros are cross-platform enough.

Well. If it won't bring a Unix Socket, it better bring at least one for IP based networking. Otherwise it'd either be files-based (like SQLite), or so super specific and special, I'd laugh myself out the window than giving a generalized answer.

Steer clear of utilizing features or functionalities that are specific to a particular platform. Stick to the common features and capabilities supported by your chosen DBMS. This ensures that the database remains compatible with different operating systems. Applying normalization techniques to organize data efficiently is a key step. Normalization reduces redundancy and helps maintain data consistency across platforms. By organizing data logically, you minimize potential discrepancies that can arise when working on different systems.

When thinking about the schema, I often find it helpful to think about which database migration strategy would be best to use. This is effectively a mechanism to update the schema as the application evolves. Some examples would be SQL scripts that ship with the application, Entity Framework Migrations etc.

IDMS applications can be ported to other mainframes and can be accessed from anywhere and called by any web language. Even phones can retrieve data from IDMS.

If you make a schema portable across database engines it will cease to utilize the benefits of each system. Using platform specific features shouldn't be ruled out but careful consideration should be given to when to use them.

It's worth noting that even standard data types like integers, floats, etc. can sometimes behave differently across platforms or database engines. One way to ensure cross-platform compatibility is to use an Object-Relational Mapping (ORM) tool or a database abstraction layer. This can help in the process of writing code that is decoupled from any specific database engine, making it easier to switch or support multiple databases.

Whow AI generated high level blah blah. True, but very high level. Backups should be part of every standard Ops procedure. Whereas security should be included by design not only in the database but in the complete system. A way of assuing this can bring us to strategies like DevSecOps.

When I think about this again the AI is right on a very high level. But does this conent really help. I've seen so many cases where performance was crashed by not taking into account concurreny, data distribution and so on. 99% of the time is a failure by design that becomes hard to correct if you do testing and optimizing as a last step. This should also be an integral part of doing this during your development and architecture phase.

Performance shouldn't be tied to the platform, always ensure best practices and monitor slow query logs for potential issues.

+1 to others This article does not resonates well with me. It isn't a good idea to run cross platform databases as it can be a easy mess. Historically (and currently) databases have been #1 single point of failures and require a good care. Think about hacker locking you out of your database etc. So moving to public cloud managed DB offerings should be right thing to promote.

The whole article is not very helpful. What is the business case here? What is the problem to be solved? Many things are true on a birds - no it's more like a looking from the moon perspective. Databases are important for the whole system design. But I don't think that focusing on cross functionality will help here. Most of the common db systems provide APIs for a varitiy of platforms and languages anyway. I've seen so many databases that were designed badly, where load scenarios, volume and database queries were not taken into account and stupid queries (that turn databases inside out mutliple times).

In cross-platform applications, utilizing REST APIs allows seamless access to data across different platforms without being constrained by the underlying database type. By abstracting the database through APIs, developers can focus on building the application without worrying about database-specific implementations, ensuring scalability and flexibility

I think this is a terrible idea and it should never be done, however if you are ever in a situation where this and it has nothing to do with a warehouse or a lake, communicate clearly how expensive and futile this exercise is going to be.