Bob Zurek

Connecting R to the Infobright® analytic database offers a powerful combined solution for handling deep analysis and statistical computations over the data contained in an Infobright database.

Click here to learn more about how the two work together.

Using R with Infobright

While query performance is one important measure of the value of an analytic database, it is not the only one that matters. Successful IT professionals and decision makers look beyond benchmarks to assess:

- Depth and breath of the feature set including but not limited to such capabilities as native operating system support, functional coverage, data type coverage, SQL syntax coverage, internationalization support

- Ease of installation and management

- Hardware and software footprint, including storage requirements 

- Software quality and stability

Beyond technical capabilities and features, one must also look at the company in terms of:

- Number of successful production references

- Track record for delivering expanded capabilities in short periods of time

- Customer and community responsiveness and support 

- Strength of both the field engineering team and professional services team 

- Company track record including strength of the management team and the investors behind the company

When I was an analyst at Forrester Research, I always told companies to ignore or look beyond competitive benchmark data or their marketing department’s comparative matrix as anyone can create a benchmark or matrix that shows their strengths over perceived weaknesses of a competitive product. I think you will hear this from most analysts in the market. These analysts will point out that companies will always try to build a competitive matrix that shows their competitive advantages over their competitors, and that you should be suspect. 

What is more useful to IT professionals is learning about how other customers have used a given product in production, the results they have had, and the effort it took to be successful. In the case of database, assess whether it can deliver the benefits you are looking for, based on your particular application, quickly and without unnecessary effort, and deliver a strong ROI from the investment in the software. 

A recent star-schema based benchmark on the MySQL Performance Blog illustrates the importance of looking beyond one set of numbers. Many applications for which companies turn to columnar databases do not use star schemas. This is certainly the case for Infobright. Extrapolating from one benchmark to make general assumptions about the performance of any particular application (of a star-schema application or not) will often lead to incorrect conclusions. With that in mind, these benchmarks cannot and should not be used as the only vehicle to show competitive performance advantages in real world customer scenarios nor be the basis for choosing one database over another.  As I pointed out above, there are considerations beyond performance that must be considered in the assessment of any database. 

We live in a competitive world and one way some companies try to compete is to throw all kinds of "benchmarks" to attempt to sway you in their direction. When you see this, be suspect. When evaluating an analytic database, ask the other key questions:

- How many customers do you have? How many successful production deployments?

- How much hardware do I need, both servers and storage?

- What is my overall TCO going to be, considering both capital and operational costs?

- What is your track record in delivering a highly stable, reliable product?

- What is the breadth of features and operating environments you support?

- What is the track record of the company and likelihood you will be around in five years?

Ferrari is known for speed and they don't have to put out benchmarks. They win races. We all know, when it comes to Ferrari, it's more than just speed that has gotten them to where they are in the market.