paul vidal - pragmatic big data nerd

Choosing the correct DBMS – Gartner Reprint review

by paul 0 Comments
Choosing the correct DBMS – Gartner Reprint review
What do you mean my speed isn't linearly scalable?

If you haven’t had the chance to look at it yet, I encourage you to read this gartner reprint: Critical Capabilities for Operational Database Management Systems. This report is extremely interesting if you’re into data at all. I’m in too deep, so I’m going to talk about this report for the whole length of this marvelous post.

The importance of specific use-cases

The first thing that jumped out to me, was that Gartner uses the word DBMS Gartner, thus highlights the fact that the dichotomy between traditional relational database management systems and what has been labelled “NoSQL” is fading out. Instead, Gartner advises to “Classify the use cases under consideration and map them to the costs, deployment options and skills requirements of the products evaluated here.”. This is extremely important and a departure from some of the preconceptions I witness amongst my fellow professionals. Often enough, I get confronted with consultants trying to categorize DBMS by capabilities (distribution capabilities, support of languages, etc.). More importantly, these platforms are marketed through those capabilities. As I argued before, and confirmed by this report, end users, consultants and software provides alike should select, recommend and market according to the use-cases at which they excel instead of the capabilities inherent to a specific platform (see previous post: The importance of specialization in software sales).

Evaluation criteria

But enough surrendering to my own confirmation biases in order to pat myself on the back and delude myself into thinking that my observations may be going in the same direction as Gartner’s. The report evaluates different vendors (selected themselves base on the set of Inclusion Criteria), using the following criteria:

  • High-Speed Ingest and Processing
  • ACID Support
  • Tunable Consistency
  • Multimodel Support
  • Automated Data Distribution
  • Cloud/Hybrid Deployment
  • Programmability for HTAP
  • Administration and Management
  • Security
  • These criteria are then weighted according to four different use cases:

  • Traditional Transactions
  • Distributed Variable Data
  • Lightweight Events and Observations
  • Hybrid Transactional/Analytical Processing (HTAP)
  • I’m not going to spend time describing the criteria, Gartner put up very readable charts to compare the different vendors. In short, it seems that Oracle is leading the traditional transaction world while DataStax is leading the distributed one. On a personal note, I’m super excited for DataStax, I get to work with many members of their team and the company I am working for leverages their solution, so it’s excellent recognition.

    I would however perhaps have had another two criteria: integration ecosystem and cost. Regarding the latter, I would have created two sets of charts: one considering the cost and one not considering it. Of course, I understand cost is a delicate and fluctuant subject and I understand Gartner’s decision. Integration ecosystem however is very important. Being able to evaluate how easy it is to integrate and use data once it is in these DBMS is extremely important when making an architecture choice.

    Personal Conclusion

    I’m always impressed by the conciseness of Gartner reports. This one does not fail in that regard, and gives a very good basis to one evaluating data management systems. That being said, and to make sure that horse is dead, think of your use-case before going for an RFP. Many DBMS can do many things, but few excel at all use-cases.

    Leave a reply

    Your email address will not be published.

    You may use these HTML tags and attributes:

    <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>