You Just Want a Humvee of a Database

May 13, 2014

The rise and fall of the Humvee have a few lessons for us in the database industry. Short of the artillery, some of the coolest features of the Humvee developed for military purposes—because the old Jeeps just couldn’t cut it—were incorporated into consumer vehicles.

Humvees were positioned as family vehicles, and television commercials showed soccer moms cruising around town in them. They popped up everywhere on the road in all colors. But the excitement didn’t last long, as people realized they were difficult to maneuver through the supermarket parking lot, expensive to maintain and the gas mileage was terrible.

How many do you see on the road today?

NoSQL solutions are the Humvees of the database industry. They are specialized, they are powerful, they are the ‘latest thing’. Like the original Humvees designed to tackle difficult terrain, NoSQL solutions target tough problems. And like the consumer version of the Humvee, NoSQL solutions are being positioned as the answer to a range of evolving data challenges. However, these new data questions can just as easily be addressed by the database already parked in the driveway you call your data center.

Postgres, for example, is the relational database with both longstanding and new capabilities that can support virtually all of today’s data types as well as unstructured and semi-structured data. This means Postgres can power many applications written for NoSQL technologies, which were developed to address so-called Big Data challenges. It also means that developers can build applications with Postgres that achieve the same result as NoSQL solutions.

Comfort and Styling—and Power

The regiment of NoSQL converts would have us believe the relational database just can’t cut it anymore and should go the way of the old Army Jeep that the Humvee replaced. But just like any single-purpose tool, NoSQL technologies cannot support all of the needs of today’s enterprises. They lack ACID compliance or even basic relational qualities for company data—they have flexibility but little durability. Pity the poor DBAs posting to online forums questions on how to create joins with their new NoSQL solution. “Sounds like need a relational database,” is the most common reply.

What’s more, with JSON and HStore, Postgres can support applications that require a great deal of flexibility in the data model. With JSON functions, Postgres can function as a document database. And HStore, which enables users to build a key/value store in Postgres, allows database administrators to store very different types of records with different attributes in the same table and query them using SQL. It’s a particularly handy tool for web developers or someone building an application that requires the ACID properties of Postgres and NoSQL capabilities. In fact, HStore pre-dates many NoSQL database projects, having been introduced in PostgreSQL v8.2 in 2006. Its popularity has expanded in recent years with new demands for working with semi-structured data.

By exploring the expanding capabilities of the relational database already in the driveway, database professionals will find they can solve new problems with what they’re already familiar with–Postgres.

To learn more about this, check out our webinar, Postgres Advances on NoSQL, on Thursday, May 15 at 11 am EDT. Presenting are Bruce Momjian, senior database architect who co-founded the PostgreSQL Global Development Community, and Marc Linster, senior vice president of products and services.


Share this