Over a million developers have joined DZone.

The Hidden Costs of Half a Database

DZone 's Guide to

The Hidden Costs of Half a Database

The biggest costs are the ones you don’t see. The best solutions are the ones that guarantee you never will.

· Database Zone ·
Free Resource

It costs plenty to install, secure, and implement a good database solution. What balloons the costs is when you need additional plugins to make the database meet your business needs. Even if the database itself is within your budget, you have to factor in buying new hardware, diverting additional developers who know the plugin technology, and of course, paying the database provider to provide “support” in integrating all these separate parts.

On top of all this mess, these plugins add additional operational layers to your data, killing your performance.

Preventing Hidden Costs

These are the three pillars that every respectable database should stand on.

  • Data Storage
  • MapReduce for sorting and searching
  • Batch Processing

Some of the bigger databases out there have one, but not all three. They offer state of the art data storage, but in order to perform MapReduce, you need an ad hoc solution. Others offer MapReduce, but in order to perform Batch Processing, you need a third-party solution like Hadoop.

This will cost you:

  • It takes longer to set up. Having to set up additional plugins means more time to launch your application. It also means more testing, coding, and overall work to release new versions, stretching your release cycle.
  • You need more people. Not every developer knows the ins and outs of Hadoop, Elasticsearch, the database you are using, and other plugins you need to make it all work. More developers have to be put on the project.
  • You need more hardware. Some of these solutions require lots of additional servers, machines, and electricity, which costs.
  • Performance issues. Hadoop is its own world. If you are storing data on your database and you need to process, Hadoop will take the data, do it’s magic, put it back to the database, and then send it to you. There are no on the fly updates to the data itself. This additional layering will cost you in speed. Additional plugins further add layers, hampering your overall performance.
  • Support issues and related costs. The real killer is when you need help putting it all together. The one place with the most experience is the database company itself. They charge for support. They charge for consulting. With so many moving parts, something is bound to unhinge from time to time. They take home lots of your money when it’s complicated, giving them every reason to keep it scattered.

The best databases, like RavenDB, integrate all of this into a complete and cohesive solution. To save you time and money, they have built into their systems MapReduce, batch processing, and even setup wizards that turn a process that can take days or weeks into minutes. In doing so, you don’t need a team of developers who need to know multiple systems. You don’t need extended tech support to screw all the pieces together. You just need one person who knows one database.

The biggest costs are the ones you don’t see. The best solutions are the ones that guarantee you never will.

database ,raven db ,hadoop ,mapreduce ,batch processing ,data storage

Opinions expressed by DZone contributors are their own.

{{ parent.title || parent.header.title}}

{{ parent.tldr }}

{{ parent.urlSource.name }}