Over the past several years there has been an enormous amount of press over NoSQL, which is a way to store and retrieve data found in a less structured and less consistent form than relational databases. NoSQL was claimed to be the heir apparent to the relational database, the kind that made Oracle an enormous company and that has been the world’s standard for decades. Keep in mind that the success of the relational database allowed Larry Ellison to buy the Hawaiian Island of Lanai. Disrupting something that big would be noticeable, but to the contrary, the stories about NoSQL have become fewer and farther between.
NoSQL was held up to the be the way to horizontally scale Big Data and Web applications and to bring a new level of simplicity to data storage and retrieval. So what happened to the buzz around NoSQL?
The NoSQL Challenges
Here are some of the challenges that make NoSQL’s adoption slower than expected:
- Transactional applications are still well-suited to relational databases with all of its built-in functionality that keep transactions intact, even if the power goes off.
- NoSQL databases are silos of their own and a challenge to move data between. Until better standards arrive, and that may be the JSON, document-based NoSQL, there’s risk of data portability with NoSQL.
- If your data isn’t enormous, moving off a relational database isn’t a necessity. People tend to postpone big changes that aren’t a necessity.
But on the Other Hand…
Then again, there are reasons to consider NoSQL despite the cons of NoSQL:
- The price structure of the relational database king, Oracle. Their customers are unhappy and stuck.
- Data will continue to grow and become a bigger problem to solve, opening more doors for NoSQL.
- Unstructured data is being vacuumed up at a remarkable rate. If you’re not sure what data you’ll need, better to put it into a NoSQL store than to ignore it.
- Managing complex data structures is easier from a design and coding standpoint with NoSQL.
So why the lower buzz around NoSQL? When I asked my friend Piyush Govil, CEO of Fabless Labs, he said, “SQL has such a well-understood way of storing and querying data that it becomes hard to replace until something better comes along and is equally understood. Also, Pig and Hive are like SQL and can be used to store and retrieve data from a Hadoop query, keeping us from having to use NoSQL when we’re working with Hadoop.”
There you have it. It isn’t that NoSQL isn’t an important advance in information technology … it just isn’t the revolution that the past buzz would have indicated.