TL;DR: not cost of license, but better software, better support, easier integration, long-term control, and innovation.
While the cost of licenses is what a lot of people focus on, that isn't the real value of open source databases like MariaDB/MySQL, PostgreSQL, ArangoDB, Mongo, H2, Cassandra, and pretty much all of the Hadoop Ecosystems. Even as exorbitant as some DB server licenses are (esp. when you do that comparison that is prohibited in your TOS for you to even think about comparing them or benchmarking the software you just paid for) the big cost is people time. The ready availability of on-line training, blogs, books, etc. make it pretty easy to have people become experts at use.
Some of the other benefits are less easily appreciated, but very real.
Product quality in main open source databases is exceptional, and patches to major problems usually come out immediately. Companies who are hyper-paranoid about security can have the source code audited, and build the software from scratch with total control. Some time ago, the DoD, when they required even commercial software to submit source code, found that open source RDBMS like MySQL and PostgreSQL had about 1/5 as many bugs or security problems. Now, most of these are never manifest, but still, Torvaldes' adage that "given enough eyes all bugs are shallow" has some truth.
The other is support. While CIOs love to be able to blame a vendor, getting things fixed is a higher priority for the rest of us. If you cannot find the answer w/ a quick search of Google or ChatGPT, posting to the project listserv usually gets you a half dozen solution pretty quickly (followed by a long debate over which is betters, someone telling you that their software project would have been a better choice in the first place, oy vey, OK, so not without its "features"). I have submitted thousands of bugs, crashes, and freezes to Microsoft and never heard from them again. Often the problem persisted even after updates or upgrades.
With open source projects, I usually get a request for additional information from the team, as well as being notified when they work on the bug and when it is fixed--and quite often have had support directly from the author of the code that was misbehaving. I cannot count the number of times I have reported a bug with some open source project and just seen it get addressed in hours to days. With one exception (KDE has been using the time format based on the C++ library rather than letting the user specify how they want it done. I am told it is a lot of work to fix, but they will get to it) crap gets fixed quickly in open source, even huge projects like KDE, MariaDB, or Eclipse. You just don't get service like this from a big commercial vendor.
Some commercial products have developed good user communities of mutual support (once you find them--just don't expect to see any helpful hacks of iOS on Apple.com). This has, AFAIK, always been a feature of open source projects. The level of technical expertise I get for free for open source software far exceeds anything I have ever received from the largest software companies. Plus, if you want, people who tell you how to fix the problem are very good about explaining why--not a feature of customer support who wants you off their line as quick as can be, even if they have to tell you the wrong solution rather than mess up their review analysis by handing you off to someone who can help.
Fellow users help you because they want people to like the software they like, and thrive on the thanks and recognition they get from others for difficult solutions or putting in extra effort to explain. This was something very rare before with commercial software (and I am sure that some licenses from a certain Bay Area DB vendor probably had a clause that somehow made it sound like you couldn't form a user group--or at least that is what some corporate lawyers have told their employees who asked if others were having a similar problem online.).
Moreover, there is better standards adherence. Open source enterprise software lives or dies based on how well it can integrate. Often the design of the open source project is predicated on the standard, as it is always easier to implement a feature that is actually documented.
Typically, there is a wide range of other software which can interface or interoperate with open source. The only database software I have ever used that didn't have MySQL/MariaDB, PostgreSQL, Firebird support out of the box was SQLyog, which is a great Windows (but runs fine on Linux with WINE) MySQL/MariaDB only GUI front end. (it has a hobbled community version for free, but it lacks code completion. Fortunately, it isn't expensive. If you use Windows and are only using MariaDB or MySQL you probably want to drop $100 on a license. Unfortunately, they don't support or even admit to plans to support PostgreSQL
Finally, innovation is clearly seen in open source projects, which tend to grow and evolve very quickly. The more I learn about what PostgreSQL is capable of I am just in awe. Similarly, the more I learn about how Linux works, the degree of cleverness and elegance in the design of things stands out, esp. when compared to commercial alternatives.
PeerSpot users agreed that functionality is of utmost importance to a quality Open Source Database (OSD). The specifications will change depending on the task you are trying to accomplish, but any Open Source Database needs to be solidly functional or there is nothing to work with. On an individual basis, scalability, metrics, and security are important features to look for. Users were clear that the efficiency of the medium which will connect the OSD with the application running it is...
TL;DR: not cost of license, but better software, better support, easier integration, long-term control, and innovation.
While the cost of licenses is what a lot of people focus on, that isn't the real value of open source databases like MariaDB/MySQL, PostgreSQL, ArangoDB, Mongo, H2, Cassandra, and pretty much all of the Hadoop Ecosystems. Even as exorbitant as some DB server licenses are (esp. when you do that comparison that is prohibited in your TOS for you to even think about comparing them or benchmarking the software you just paid for) the big cost is people time. The ready availability of on-line training, blogs, books, etc. make it pretty easy to have people become experts at use.
Some of the other benefits are less easily appreciated, but very real.
Product quality in main open source databases is exceptional, and patches to major problems usually come out immediately. Companies who are hyper-paranoid about security can have the source code audited, and build the software from scratch with total control. Some time ago, the DoD, when they required even commercial software to submit source code, found that open source RDBMS like MySQL and PostgreSQL had about 1/5 as many bugs or security problems. Now, most of these are never manifest, but still, Torvaldes' adage that "given enough eyes all bugs are shallow" has some truth.
The other is support. While CIOs love to be able to blame a vendor, getting things fixed is a higher priority for the rest of us. If you cannot find the answer w/ a quick search of Google or ChatGPT, posting to the project listserv usually gets you a half dozen solution pretty quickly (followed by a long debate over which is betters, someone telling you that their software project would have been a better choice in the first place, oy vey, OK, so not without its "features"). I have submitted thousands of bugs, crashes, and freezes to Microsoft and never heard from them again. Often the problem persisted even after updates or upgrades.
With open source projects, I usually get a request for additional information from the team, as well as being notified when they work on the bug and when it is fixed--and quite often have had support directly from the author of the code that was misbehaving. I cannot count the number of times I have reported a bug with some open source project and just seen it get addressed in hours to days. With one exception (KDE has been using the time format based on the C++ library rather than letting the user specify how they want it done. I am told it is a lot of work to fix, but they will get to it) crap gets fixed quickly in open source, even huge projects like KDE, MariaDB, or Eclipse. You just don't get service like this from a big commercial vendor.
Some commercial products have developed good user communities of mutual support (once you find them--just don't expect to see any helpful hacks of iOS on Apple.com). This has, AFAIK, always been a feature of open source projects. The level of technical expertise I get for free for open source software far exceeds anything I have ever received from the largest software companies. Plus, if you want, people who tell you how to fix the problem are very good about explaining why--not a feature of customer support who wants you off their line as quick as can be, even if they have to tell you the wrong solution rather than mess up their review analysis by handing you off to someone who can help.
Fellow users help you because they want people to like the software they like, and thrive on the thanks and recognition they get from others for difficult solutions or putting in extra effort to explain. This was something very rare before with commercial software (and I am sure that some licenses from a certain Bay Area DB vendor probably had a clause that somehow made it sound like you couldn't form a user group--or at least that is what some corporate lawyers have told their employees who asked if others were having a similar problem online.).
Moreover, there is better standards adherence. Open source enterprise software lives or dies based on how well it can integrate. Often the design of the open source project is predicated on the standard, as it is always easier to implement a feature that is actually documented.
Typically, there is a wide range of other software which can interface or interoperate with open source. The only database software I have ever used that didn't have MySQL/MariaDB, PostgreSQL, Firebird support out of the box was SQLyog, which is a great Windows (but runs fine on Linux with WINE) MySQL/MariaDB only GUI front end. (it has a hobbled community version for free, but it lacks code completion. Fortunately, it isn't expensive. If you use Windows and are only using MariaDB or MySQL you probably want to drop $100 on a license. Unfortunately, they don't support or even admit to plans to support PostgreSQL
Finally, innovation is clearly seen in open source projects, which tend to grow and evolve very quickly. The more I learn about what PostgreSQL is capable of I am just in awe. Similarly, the more I learn about how Linux works, the degree of cleverness and elegance in the design of things stands out, esp. when compared to commercial alternatives.