For three decades, SQL has been the universal data research tool. It is flexible, powerful and, most of ,all universal. Business users who master it can work with virtually any data in any enterprise. It lets the user define what needs to be done with what data and leaves it to the database to find that data and how to best get the job done.

But SQL and the databases that use it are designed for an earlier age, where data meant structured-only and where development meant “waterfall,” where new versions of an application took a year or more to produce and where changes to the database were major projects involving months of work by teams of developers. SQL databases are the province of the IT organization (ITO), which imposes strict rules and processes designed to maximize control and, in the process, minimize  agility, writes Wikibon Analyst George Gilbert in the third part of his series on how Systems of Intelligence drive database proliferation.

For instance, adding a column to an SQL database requires a process that starts with sending a request to the service desk and waiting for a ticket back from IT. The developer then may get a new ticket asking a clarifying question or two. When those are answered, storage admins have to review the space impact, security admins have to assess the impact on access control and database admins have to determine the performance impact on existing queries, all before the code writing can start. Then coding with SQL requires developers who know the language and who can constantly shift between it and whatever language they are using to program..

This does not fit the new, highly iterative, DevOps approach to developing applications, where change is a given and agility is a necessity. Especially when developing user-facing applications and Systems of Intelligence, the waterfall approach is just too slow. That is a major reason why NoSQL databases in general, and MongoDB in particular, have become popular with developers working on new-style applications.

The trade off, Gilbert says, is that it becomes harder to bring all the data in the organization together for analysis. The advantage is that the organization can keep up with the very rapid pace of technological development that marks this new century.

All three of this series, along with other pieces of George Gilbert’s analysis of the cloud environment, are available on the Wikibon Premium website.

Image by Nowman via Pixabay