In big corporations you don't get fired for buying Oracle. BTW Oracle bought Sun Microsystems because it was the most widely used hardware platform for Oracle. On that basis Oracle could be considered the king of RDBMS. But it is nothing special technically and it is very expensive to purchase and the support contracts are eye watering. Larry Ellison can afford to lavishly play at Admirals Cup on the licence income.
Several years ago a package was released that allows you to run apps that expect Oracle on Postgres.
Of course startups do not use Oracle. In fact few even use Windows. Why spend huge amounts of money when you can get better solutions for free? And there are so many good choices, including noSQL solutions.
Over time this didn't matter as many RDBMS offered advanced MVCC features and transactional guarantees "on the cheap". (MS SQL and pgsql being late bloomers to this club - there are many other good databases that do similar).
Your professor is probably a veteran from the heydays of Oracle's dominance. The meme of Oracle (i.e. if the system is developed using Oracle being default good) still carries in the minds of many financial execs and CTOs. But it is fading fast.
I personally wouldn't touch Oracle for my enterprise startup which provides BI solutions for asset managers. We use SQL Server and pgsql.
Your professor is correct in that it's widely used amongst such companies, though other competitors are very often in place as well such as Sybase or Microsoft SQL Server.
The right tool for the right job (or in the startup world, the tool you can afford.)
https://news.ycombinator.com/item?id=13726012
Oracle DB has some of the best tech. For e.g Partitioning of tables works great. DB can be seriously tuned and is a workhorse for many different use cases. They have easy access to use tools that make management super simple.
Though you pay a hefty price for all of this, often you can get similar things done with open source with a bit more hacking and deep diving though you would need to have much more technical skills than your typical oracle DBA.
Then you have what I call the 'propriety DB' tax. Which usually means your open source tools typically support oracle DB as an afterthought. I often come across an open source solution to solve some problem only to realize we need to get data in or out of oracle with support either non existent or patchy at best. PGSql/MySQL are mostly supported first.
For e.g
- I recall getting cx_oracle(python DB API) was a serious pain in the ass.
- Want to use a heroku build pack for your app to deploy to an open source paas solution? Need to roll your own custom build pack
- Need continuous integration, auto builds? Need to hack around their stupid license agreement
- Need a local dev db? You get a seriously limited version (XE) without feature parity to what is most likely running in production. You have to write specific code for local vs production in some cases - like accessing partitions etc
Really I would be surprised if there were Serious SAAS startups that were using it.
That $ could be invested elsewhere and open source is good enough.
If you're a start-up going after B2B market, and you want to offer on-prem service, I suggest you get into the Oracle developer network and try as hard as you can to make sure your product works with Oracle as a backend.
Just having it as an option for big corporations is worth the effort... But not as a first time option (unless ORA funds you, which is another story).