I used MySQL on Ubuntu for temporary storage of financial data utilized for analysis of creditworthiness as the main use case, as I had some streaming services that pulled in data from Equifax.

It helped streamline my data management processes because I needed somewhere to store the data rather than just an array in memory. I needed to have the data to have the record because at some points I have to go back to determine what was the basis of this credit assessment. The data was there and could recreate the calculation once the data was called, and the data was called timestamped, secure, and always available, which is the whole point of MySQL on Ubuntu. Otherwise, if I used an array, it would only be there for the life of the system being up; there would not be a necessity of saving any of the data that was just in a temporary array. By putting it in MySQL on Ubuntu, even if the node went down, the database would come back up.

Integration is always important regarding operating systems and these types of products, so being able to integrate and export or import from JSON structures is very critical. Sometimes that is a little complicated because of the sometimes hierarchical nature of JSON or XML data formats, which do not always match to how you can structure MySQL on Ubuntu as a third normal form. There are those sorts of things that sometimes get inexperienced people; it does not seem to make sense.

For denormalization, if you are trying to analyze it only, there is probably a shortcut that I have seen in some tools that once you define the third normal form type of data, it kind of automatically comes up with a way of analyzing it, turning it into an automated pivot table without you having to design the pivot table. Those things would be good to get the analysis.

Some of the analysis that I had to code from scratch in Python were really simple binomial algorithmic comparisons. Some of that could turn into AI functions. Instead of coding it directly, I could use normal language saying I want to analyze this data based on whether this company has good financial viability to extend a million dollars of credit for buying fuel around the world or whatever the parameter is. That is what I can see coming in the future, that somebody that does not know how to code or does not really want to spend the time coding could actually ask in natural language AI to come up with that. To some extent, I have done that more recently with ChatGPT anyway to come up with a piece of code that at the moment does not work perfectly, but it is still Python and gives me the basic framework to then make it work elegantly.

I dealt with MySQL on Ubuntu for about two years.

I assess the transactional support features of MySQL on Ubuntu as just very simple insert and read because there was only really one stream that inserted data, so there was not any multi-concurrency of entering records to the database. The multi-concurrent users were just accessing and running the algorithm in the nodes to actually get an evaluation. Basically, they called a thing, they said I wanted to give some credit to company A. The node would do a query to Equifax and Experian to get whatever they could get on that and some Dun & Bradstreet information as well, put that in the Ubuntu SQL database, MySQL on Ubuntu database, and then run another algorithm to determine based on a couple of statistical points of view whether to give them this credit or whether they have to prepay for anything.

Regarding MySQL on Ubuntu scalability, I never touched it in terms of scalability because we were not looking at terabytes of data; we were looking at gigabytes of data. I think the database hardly went above one gigabyte when I was there because it is very simple. It just says here is the name of the company and here is anything we have got from three of the main credit or fact information sources globally that might have information on that database. Then a quick search in a virtual web environment to see whether there was any more generic business information on those companies, such as whether director A has just been fired or director B has been in jail for fraud or something, to get a little sentiment analysis of all these other things. The total data was very little.

The initial setup was very straightforward because I have a lot of experience in various database technologies and in Python and creating servers in virtual servers in AWS.

MySQL on Ubuntu is very simple, easy, and quick to use for people with database expertise. For that light use of MySQL on Ubuntu, that was all I needed, so there was nothing that was inadequate, and I could easily access it. The node was fine and the accessibility for people around the world that were actually asking whether they could give credit to this company or whether they have to pay up front was the main thing that was being supported, so everything was fine. There were no limitations there because the data volume was fairly small; globally, they only probably looked at about a thousand different entities per month and so there was only about another thousand records each month added, and the analysis done to give a pretty much real-time determination of whether to extend credit or request prepayment.

AWS was basically my main cloud provider with this. The low cost is what I liked about using MySQL on Ubuntu because basically, I did not have much of a budget for the solution, just my time and a few units of AWS services to work on, because it had to be more than just something on my own PC in the office, so other people could access it, allowing me to actually create a front end as well with it. It is very lightweight regarding the pricing; I never got any issues and was within my department budget for all AWS services for development. We never actually got a production budget for it because things were changing and then COVID hit as well, so it slowed down the demand. I am not quite sure what they did with that solution after that company, but I know they were using it. I still sometimes get an error message that somehow gets into my current AWS account.

I just utilized the standard virtual high availability options on Ubuntu, so I had redundant nodes in two regions. I dealt with MySQL on Ubuntu a little bit, but we never really got the Docker setup completed; I had some experience working with it. I have still maintained some Redshift analysis and some code in Python on some AWS products in the last twelve months. I am not working day-to-day anymore in that area with the Amazon solutions by chance. I deal with a little Amazon Linux and maybe Elastic Disaster Recovery, but not in detail, so I am probably not really the best candidate at the moment. I rate MySQL on Ubuntu a ten out of ten.