OK, I know. ”Big Data”. It’s “big”. It’s “in”. It’s important. It’s true. It’s here to stay.
Now what? First, some stats:
1. 90% of the world’s data was produced in the last two years.
2. The NoSQL database market is projected to be about $3.5 billion by 2018.
3. The RDBMS market, according to Gartner, is currently a $26 billion market and is growing about 9% per year. (That puts the market at $40 billion by 2018.)
Now, some observations:
1. 42% of the world’s data volume seems to be either: (a) videos of cats dancing (b) pictures of legs and feet with a beach background (c) someone else’s amazing kids’ video of some skill they have, or (d) well, you get the picture.
2. There may already be more NoSQL/Big Data database projects/companies than there are SQL database companies.
3. Most enterprises are not Google, Yahoo! or Facebook.
4. Sometimes, when you’re looking for a needle in a haystack, you only find a stack of hay.
My point is this: yes, big data is important, and yes, with the proper infrastructure, tools and resources, one may be able to extract value from all that data, but do not, for a moment, forget the continued importance of your “non-sexy,” “boring,” back-office, transactional applications, for which your RDBMS is loyally, dependably, securely and efficiently inserting, updating, storing and protecting your company’s most important data.
The problem with traditional relational databases, I’d suggest, actually has to do with the fact that they are housing our most important data. That’s exactly why they’re so expensive. Companies like Oracle, SAP, IBM and Microsoft know that, and they also know that since basically every database uses its own SQL, it’s difficult to move from one to another. The result, for example, is Oracle’s $47,500 license fee per core, plus extras (replication, partitioning, etc.).
So, does that mean we should all abandon relational database for these “new-fangled shiny red buttons”? Of course not.
Every so often, some new database technology is introduced as “the next big thing” that will unseat RDBMSs. This happened in the 90′s with object oriented databases. Then, at the turn of the century (I love saying that), XML storage was all the rage. Now, it’s document storage, key value stores, non-durable storage, etc.
But, here’s what happens…Database companies review and evaluate the new technology, determine its value and viability, and then integrate it into their relational database product, giving the customer the best of all worlds.
PostgreSQL has done this, too, and with the release of 9.3 there are more features than ever, with better performance than ever, that allow customers to standardize on a great database foundation. What’s more, PostgreSQL doesn’t come with the sticker price the other big-time proprietary products come with, which allows us to re-direct that spend to other initiatives, like finding that needle.Tweet