This post is an extension of last week’s post, intended to be a series based on career highlights with interludes on my latest challenge in completing my Master’s in Data Analytics. If something doesn’t make sense, look to the previous blog entry; I will link and date them moving forward for clarity.

Throughout my career, I’ve advocated for open-source software when it was appropriate to the situation.

For example, after a particularly grueling audit, one that could have been avoided, the head of one organization asked me to explore alternatives to their existing Oracle stack.

A subset of the management team had:

• Licensed Oracle BI for a physical system, but propogated it to multiple virtualized environments.
• Oracle’s BI was then rebranded, marketed, and sold as the organization’s reporting tool.

NOTE: I won’t go into specifics on why these are monumental issues, as I am well aware; if the reader is not, they should feel free to research them and their legal ramifications. It should be obvious, especially for those of us who actively managed and facilitated software licensing.

This successfully worked to obtain Oracle’s attention, they noticed the rebranding on the organizations website, resulting in a full licensing audit, which found that from the database side, in addition to the BI problems, a licensing reseller had set them up with licensing by user for their development regions (to be cheaper) and licensing by the core for production only, Oracle did not approve of this, and it was not their practice to allow such licensing arrangements for development environments.

I knew none of this going into the audit, but I had inquired about the BI deployment and use case multiple times, only to be rebuffed; it was his project, and none of my business when I asked the project owner. I had suspicions, but what could I do, except focus on things I could do and ignore the noise. This would be the same person who would inform me that they “were going to fire me the first chance they get”, but not without first surreptitiously looking around to make sure nobody else was within earshot. What was the game here? (rhetorical question)

I could call this post, “How I Ended Up the Director of Database Architecture”, but it was not just this audit that put me there. It was also the result of productivity, ability, knowledge, improvements, and long hours of hard work that made me a candidate in the first place. Anyhow, that’s another story, so I will end this shameless plug and get back to the story at hand.

I evaluated comparable solutions from Microsoft, Sybase, and others. While each vendor calculated licensing costs differently, the totals landed in the same ballpark.

This organization operated in the financial sector, requiring encryption for data at rest and in transit. At the time, open-source alternatives couldn’t seamlessly integrate with the proprietary code already in place.

The challenge was years of code built around Oracle’s ability to transparently handle encryption for data at rest. Porting that code to a new backend without equivalent capabilities would have been arduous, not to mention the resource impact of encrypting and decrypting data on the fly every time it was read or written. Not a good fit for a high-transaction system.

However, in other areas, an open-source database like MySQL would have been a perfect fit. The application in question was write-heavy, with straightforward transaction handling focused on processing and approving online credit applications.

In October 2015, MariaDB, a branch of MySQL, announced support for transparent data encryption in version 10.1. Having already completed evaluations of MySQL, this instantly made MariaDB a viable contender.

Speed was relative. If the application had required heavy reads and writes involving large complex datasets and objects, performance concerns would have been more pressing.

Many open-source solutions, and their proponents, tout performance claims that rival or exceed Oracle. But while speed is relevant, it should take a back seat to the specific methods and data used to achieve those benchmarks.

In reality, these results are often achieved by optimizing hardware, leveraging distributed architectures, and tailoring workloads, all while processing relatively simple transactions. This critical context is frequently overlooked, which is why some adopted open-source solutions end up delayed or fail altogether.

The solution must fit the scenario. Flashy marketing claims based on narrowly scoped, highly controlled performance tests are rarely relevant in the real world, yet there is a cultish following that takes none of this into consideration.

Oracle still dominates in the realm of complex transaction handling, enterprise-grade security and compliance, and an integrated stack covering every key area for massive mission-critical systems in areas where speed, security, capability, and stability are relevant.

Which should be everywhere, right?

What do I know? Per the previous blog entry, these people have claimed I was difficult to work with, I was non-technical, claims were made that I didn’t know what I was doing, and that I was a poor manager, among other things. Knowing how to ask the hard questions, to challenge and fix broken systems, and stand where others won’t is not indicative of any of those labels, unless, of course, those flinging such hogwash are the cause of such problems.

Funny how rebranding works.

Reference Blogs:
August 3, 2025: https://modba.net/2025/08/03/the-myth-of-being-hard-to-work-with/