-
“Recently, a lot of new non-relational databases have cropped up both inside and outside the cloud. One key message this sends is, “if you want vast, on-demand scalability, you need a non-relational database”.
If that is true, then is this a sign that the once mighty relational database finally has a chink in its armor? Is this a sign that relational databases have had their day and will decline over time? In this post, we’ll look at the current trend of moving away from relational databases in certain situations and what this means for the future of the relational database.”
Notional Slurry
Pontification without all the gritty gravitas
Not that you asked, but: this is a pretty bad article, given that it’s entire premise is “It’s hard to scale an RDBMS, but key/value stores magically scale for free”. Nonsense. Key/value stores still have all the hard questions that come with scaling out — how do you partition data across servers ? How do you detect and handle machines failures ? How do you deal with data re-striping when you add new machines ? etc.
It’s true that RDBMS are not always well-suited for the brave new world, but that’s not because of scale, at least the way it’s described in this article. http://db.cs.yale.edu/hstore/vldb07hstore.pdf is a much more cogent survey of the reasons.
Not that I asked
To me the value in the article comes from the list of interesting variations. And it comes at a point in my work where I’m tempted—for very pragmatic reasons—to pick up some computational complexity-prone people and shake them until their necks snap. Though I love them as people, truly I do.
I come increasingly to hate the sin of premature optimization. And having spent a long, long time drawing very, very complicated entity relationship models, while at the same time practicing emergent design on the code of some weirdly complex projects… well, I have to question.
So: yes. It’s a slapdash article. But I hadn’t heard of Voldemort before then, and I think it’s interesting.
I can say one thing: I’m troubled more all the time when folks rely on store-bought hardwired panacea solutions in low-level algorithm designs.