Recently on an Oracle certification forum, someone posted a message directed specifically to me. He indicated that since purchasing Sun several years ago, the Oracle RDBMS is waning, and MySQL and Java are now the company’s “important” technologies. As evidence of this, he noted that the organization where he is employed uses MySQL/MariaDB, and PostgreSQL in lieu of the Oracle database. He suggested that I “move with the times” and create certification guides for these products, because they will be the important relational databases in the next 10 years.

I have seen others express similar sentiments on forums before. A number of people believe that open source databases will kill Oracle sooner or later. For this particular message I responded that Oracle would probably hang on to life for a few more years. It is worth expanding, however, on the topic of Oracle being at Death’s door. Before beginning the eulogy, a disclaimer is in order. As I started writing this, a snippet from a book by the late Terry Pratchett came to mind:

The Monks of Cool, whose tiny and exclusive monastery is hidden in a really cool and laid-back valley in the lower Ramtops, have a passing-out test for a novice. He is taken into a room full of all types of clothing and asked: Yo, my son, which of these is the most stylish thing to wear? And the correct answer is: Hey, whatever I select.

Terry Pratchett, "Lords and Ladies", 1992

As a general rule, most of the IT professionals I have worked with, if asked “Yo, my son, which of these is the best technology to use?” would give the same basic answer. The best technology to use is almost invariably the one that they know the most about, or possibly the one that they are already using. The people who suggest that open-source databases will supplant Oracle are typically professionals who use an open-source database. On the other hand ... keep in mind that the author of this article has based the past 20 years of his career working with Oracle. (Yo, my son ... )

If someone were to ask me: “What is the best relational database management system currently available?” I would reply Oracle almost before they finished the question. Given that specific question, though, I doubt anyone would answer MySQL, MariaDB, or PostgreSQL. Even if the question specifically excluded Oracle, those three are unlikely to make the top rank. The question would need to include a requirement for open source or minimal licensing cost before any of these would be considered a “better” RDBMS than Oracle. The premise of the “Open source will kill Oracle” argument is essentially that corporations will decide to give up some of the advanced features and functionality available in Oracle in order to reduce licensing costs.

I consider that to be a fairly unlikely scenario. It is roughly equivalent to saying that Mercedes and BMW will soon go out of business because Kia makes cheaper cars. In reality, customers often want or need the features of higher-priced products and are willing to pay for them. A company I worked with several years ago performed statistical analysis on medical data for large health insurers. The analysis involved hundreds of complex SQL queries and their largest customers took several days to process. Anything which made those queries run faster allowed the company to make more money. Moving to a less-capable database to save money on licensing costs would have been shooting themselves in the foot.

Beyond that, the loss of data or the interruption of processes being handled by Oracle databases or applications has the potential to cost organizations millions of dollars. Even if switching to an open-source database is feasible, switching platforms tends to be both painful and labor-intensive. Recouping the costs invested in the switch could take years.

Often when someone offers career advice about a given technology, I will run a search for jobs mentioning it in a couple of geographic areas. This provides a fast rule-of-thumb estimate on its popularity. I used my favorite job site to look for listings referencing Oracle, MySQL, MariaDB and PostgreSQL. I ran the search for my own region plus two technology hubs. Based on the results below, it is difficult to argue that a mass migration away from Oracle is underway:

In a decade, the numbers above may well be extremely different. In fact, I would be shocked if that were not the case. The players in the information technology field are constantly shifting around. When I was hired by Oracle tech support almost 20 years ago, Informix and Sybase held a reasonable amount of market share. Today they are seldom mentioned in job listings except as part of a laundry list of databases (i.e. "experience with other DBMS including Oracle, Sybase, Informix, PostgreSQL, Times Ten ... "). It is remotely possible than in 10 years the Oracle database will have lost so much market share that it is no longer relevant. That is certainly not the case today.

Even given a long-range projection, however, it is difficult to imagine open-source databases ever causing the demise of Oracle. They are neither new nor are they a transformational technology. To displace Oracle would require something better, not just something cheaper but less functional. NoSQL is one technology on the rise that is viewed as a serious competitor to traditional databases. It is not an area that I am an expert in. From what I have read, the general consensus is that it had advantages over traditional SQL databases in some areas, but disadvantages in others. If that is the case, then Oracle may lose some market share to NoSQL databases, but it would be unlikely to disappear.

Predicting the future is never a simple task. Right now, there are a huge number of jobs available for people skilled with Oracle products. I don't see that changing significantly for the next five years or so. For that matter, I don't expect a change after the next five years. I just refuse to speculate on technology trends further out than that.