4 ms·

people just like hating on successful companies, a HN special.. I ignore most those types of posts on this site because they are without any merit

toyg 1y ago
I think the Oracle hate goes a bit beyond the generic hatred for successful companies (be it Microsoft, FAANG, etc). People actually used to like the products that made these companies explode. Windows 2000 was cool; Facebook was cool; Google was cool. Whenever they stop being pure unadulterated evil for a few minutes, very quickly a lot of people are willing to forgive them and welcome them back into polite society. But as long as most geeks can remember, Oracle has never been cool. At its peak, the company enabled a world of snooty BOFH DBAs, selling unreasonably expensive products to well-oiled middle-managers. And then they started "acqui-squeezing" adjacent products, blackmailing their own customers, and suing everyone in sight. They could cure cancer tomorrow, gifting all the related IPs to the world, and most geeks would still see them as scum trying to whitewash their image - and they would probably be right. There are some great engineers in Oracle, but their management is all that is wrong with capitalism.
- kragen 1y ago
  From 01979 until about 02000, Oracle's RDBMS software was probably the best in the world, and definitely better than the free-software alternatives like Postgres. (Remember that MariaDB, then named MySQL, didn't become free software until 02000 and didn't support transactions until even later.) For several years after that, it was still the best database for some purposes. And that's true even though the Bastard Database Operators From Hell were present from the very beginning. SQL is kind of shitty, but the relational database model is so great that it makes up for it. And until the rise of entity component systems, SQLite, PostgreSQL, and MariaDB, all the decent implementations were proprietary software. Separately, although SQL's implementation of transactions is even more broken than its implementation of the relational model, transactional concurrency is also great enough to make up for SQL, and, again, all the usable implementations used to be proprietary.
  - aleph_minus_one 1y ago
    > From 01979 until about 02000, Oracle's RDBMS software was probably the best in the world, and definitely better than the free-software alternatives like Postgres. [...] For several years after that, it was still the best database for some purposes. Which RDBMS software has/have become the best in the world after that?
    - cerved 1y ago
      Postgres presumably
      - aleph_minus_one 1y ago
        As far as I am aware (I may be wrong, or things have changed in the last years) Postgres still is is worse in horizontal scaling than Oracle and Microsoft SQL Server (and likely also DB2).
        kragen 1y ago
        Maybe, but I think MariaDB beats all of them at that. Also, though, horizontal scaling is a lot less important now than it was 20 or 30 years ago. https://www.servethehome.com/2025-server-starting-point-intel-amd-nvidia-broadcom-solidigm-micron-kioxia-marvell-arista/ https://www.servethehome.com/2025-server-starting-point-inte... says AMD has 192 cores per socket, and you can get two-socket motherboards, so 384 cores total. And you can stick 12 128GiB DDR5-6000 DIMMs in it, so 1.5 tebibytes of RAM, and a single SSD is 30 terabytes, and SSDs can commit a transaction group durably in typically 0.1 milliseconds. And those 384 cores (EPYC 9005, so Zen 5c, https://www.servethehome.com/amd-epyc-9005-turin-turns-transcendent-performance-solidigm-broadcom/ https://www.servethehome.com/amd-epyc-9005-turin-turns-trans...) are 2.25GHz and typically about 2.2 instructions per clock (https://chipsandcheese.com/p/zen-5-variants-and-more-clock-for-clock https://chipsandcheese.com/p/zen-5-variants-and-more-clock-f...), and they support AVX512. As one rough estimate, 2.25GHz with AVX512 (at 1 IPC) means you can do 36 billion column-oriented 32-bit integer operations per core per second, which with 384 cores means 13 trillion 32-bit integer operations per core per second. On one server. So if you have a query that needs to do a linear scan of a column in a 13-million-row table, the query might take 300μs, but you should be able to do a million such queries per second. But normally you index your tables so that most queries don't need to do such inefficient things, so you should be able to handle many more queries per second than that! (Each socket has 12 DDR5 channels, totaling 576 gigabytes per second to DRAM per socket or 1.13 terabytes per second across the two of them, so you'll get worse performance if you're out of cache. And apparently you can use 512GiB DIMMs and get 6 tebibytes of RAM!) So, if you need more than one server for your database, it's probably because it's tens or hundreds of terabytes, or because you're handling tens of millions of queries per second, or because your database software is designed for spinning rust. Spinning rust is still the best way to store large databases, but now the cutoff for "large" is approaching the petabyte scale. I think the space of databases that are under ten terabytes and under ten million queries per second is large enough to cover almost everything that most people think of when they think of "databases".

toyg 1y ago

I think the Oracle hate goes a bit beyond the generic hatred for successful companies (be it Microsoft, FAANG, etc). People actually used to like the products that made these companies explode. Windows 2000 was cool; Facebook was cool; Google was cool. Whenever they stop being pure unadulterated evil for a few minutes, very quickly a lot of people are willing to forgive them and welcome them back into polite society. But as long as most geeks can remember, Oracle has never been cool. At its peak, the company enabled a world of snooty BOFH DBAs, selling unreasonably expensive products to well-oiled middle-managers. And then they started "acqui-squeezing" adjacent products, blackmailing their own customers, and suing everyone in sight. They could cure cancer tomorrow, gifting all the related IPs to the world, and most geeks would still see them as scum trying to whitewash their image - and they would probably be right. There are some great engineers in Oracle, but their management is all that is wrong with capitalism.

kragen 1y ago
From 01979 until about 02000, Oracle's RDBMS software was probably the best in the world, and definitely better than the free-software alternatives like Postgres. (Remember that MariaDB, then named MySQL, didn't become free software until 02000 and didn't support transactions until even later.) For several years after that, it was still the best database for some purposes. And that's true even though the Bastard Database Operators From Hell were present from the very beginning. SQL is kind of shitty, but the relational database model is so great that it makes up for it. And until the rise of entity component systems, SQLite, PostgreSQL, and MariaDB, all the decent implementations were proprietary software. Separately, although SQL's implementation of transactions is even more broken than its implementation of the relational model, transactional concurrency is also great enough to make up for SQL, and, again, all the usable implementations used to be proprietary.
- aleph_minus_one 1y ago
  > From 01979 until about 02000, Oracle's RDBMS software was probably the best in the world, and definitely better than the free-software alternatives like Postgres. [...] For several years after that, it was still the best database for some purposes. Which RDBMS software has/have become the best in the world after that?
  - cerved 1y ago
    Postgres presumably
    - aleph_minus_one 1y ago
      As far as I am aware (I may be wrong, or things have changed in the last years) Postgres still is is worse in horizontal scaling than Oracle and Microsoft SQL Server (and likely also DB2).
      - kragen 1y ago
        Maybe, but I think MariaDB beats all of them at that. Also, though, horizontal scaling is a lot less important now than it was 20 or 30 years ago. https://www.servethehome.com/2025-server-starting-point-intel-amd-nvidia-broadcom-solidigm-micron-kioxia-marvell-arista/ https://www.servethehome.com/2025-server-starting-point-inte... says AMD has 192 cores per socket, and you can get two-socket motherboards, so 384 cores total. And you can stick 12 128GiB DDR5-6000 DIMMs in it, so 1.5 tebibytes of RAM, and a single SSD is 30 terabytes, and SSDs can commit a transaction group durably in typically 0.1 milliseconds. And those 384 cores (EPYC 9005, so Zen 5c, https://www.servethehome.com/amd-epyc-9005-turin-turns-transcendent-performance-solidigm-broadcom/ https://www.servethehome.com/amd-epyc-9005-turin-turns-trans...) are 2.25GHz and typically about 2.2 instructions per clock (https://chipsandcheese.com/p/zen-5-variants-and-more-clock-for-clock https://chipsandcheese.com/p/zen-5-variants-and-more-clock-f...), and they support AVX512. As one rough estimate, 2.25GHz with AVX512 (at 1 IPC) means you can do 36 billion column-oriented 32-bit integer operations per core per second, which with 384 cores means 13 trillion 32-bit integer operations per core per second. On one server. So if you have a query that needs to do a linear scan of a column in a 13-million-row table, the query might take 300μs, but you should be able to do a million such queries per second. But normally you index your tables so that most queries don't need to do such inefficient things, so you should be able to handle many more queries per second than that! (Each socket has 12 DDR5 channels, totaling 576 gigabytes per second to DRAM per socket or 1.13 terabytes per second across the two of them, so you'll get worse performance if you're out of cache. And apparently you can use 512GiB DIMMs and get 6 tebibytes of RAM!) So, if you need more than one server for your database, it's probably because it's tens or hundreds of terabytes, or because you're handling tens of millions of queries per second, or because your database software is designed for spinning rust. Spinning rust is still the best way to store large databases, but now the cutoff for "large" is approaching the petabyte scale. I think the space of databases that are under ten terabytes and under ten million queries per second is large enough to cover almost everything that most people think of when they think of "databases".