7 hard truths about the NoSQL revolution

Forgoing features for speed has its trade-offs as these NoSQL data store shortcomings show

By Peter Wayner, InfoWorld |  Big Data, nosql

There are few extras in NoSQL databases. If you want to do anything but store and retrieve data, you're probably going to do it yourself. In many cases, you're going to do it on a different machine with a complete copy of the data. The real problem is that it can often be useful to do all of the computation on the machine holding the data because shipping the data takes time. But tough for you.

NoSQL solutions are emerging. The Map and Reduce query structure from MongoDB gives you arbitrary JavaScript structure for boiling down the data. Hadoop is a powerful mechanism for distributing computation throughout the stack of machines that also holds the data. It is a rapidly evolving structure that offers rapidly improving tools for building sophisticated analysis. It's very cool, but still new. And technically Hadoop is an entirely different buzzword than NoSQL, though the distinction between them is fading.

NoSQL hard truth No. 7: Fewer toolsSure, you can get your NoSQL stack up and running on your server. Sure, you can write your own custom code to push and pull your data from the stack. But what if you want to do more? What if you want to buy one of those fancy reporting packages? Or a graphing package? Or to download some open source tools for creating charts?

Sorry, most of the tools are written for SQL databases. If you want to generate reports, create graphs, or do something with all of the data in your NoSQL stack, you'll need to start coding. The standard tools come ready to snarf data from Oracle, Microsoft SQL, MySQL, and Postgres. Your data is in NoSQL? They're working on it.

And they'll be laboring on it for a bit. Even if they jump through all of the hoops to get up and running with one of the NoSQL databases, they'll have to start all over again from the beginning to handle the next system. There are more than 20 different NoSQL choices, all of which sport their own philosophy and their own way of working with the data. It was hard enough for the tool makers to support the idiosyncrasies and inconsistencies in SQL, but it's even more complicated to make the tools work with every NoSQL approach.

This is a problem that will slowly go away. The developers can sense the excitement in NoSQL, and they'll be modifying their tools to work with these systems, but it will take time. Maybe then they'll start on MongoDB, which won't help you because you're running Cassandra. Standards help in situations like this, and NoSQL isn't big on standards.


Originally published on InfoWorld |  Click here to read the original story.
Join us:
Facebook

Twitter

Pinterest

Tumblr

LinkedIn

Google+

Answers - Powered by ITworld

Join us:
Facebook

Twitter

Pinterest

Tumblr

LinkedIn

Google+

Ask a Question
randomness