Is MySQL Doomed to Extinction?
opensource.sys-con.com
opensource.sys-con.com
I know that it's become fashionable to write of MySQL or SQL in general and many large sites are migrating away from it to non-relational data stores. However, for many sites, MySQL does a wonderful job and relational features cut down development time. The founder of Flickr spoke to this at a Django conference saying something to the effect of, "ever since there was more than 200 sites on the internet, it has been better to target a framework at the rest of the internet than at the top 100 sites". Specifically, he was speaking to sharding and other highly scalable techniques that Django doesn't deal with natively and saying that the vast, vast majority of sites will never have to deal with those questions.
Relational technology definitely has its place and does offer developers advantages. And relational technology does scale for some high-traffic operations. Look at Wikipedia. It's the 6th most trafficked site in the world and backed by MySQL. Granted, they do use caching and other techniques as well, but it's a highly available, highly trafficked site. In fact, I can't remember the last time I heard about Wikipedia going down.
New technology directions are great. I'm really excited about Cassandra. Do we really have to declare something dead and something else the winner?
Mr. Someone, who just started a business competing with <specific technology> says yes!"
Which means you can't simply dismiss his statements; you have to look into them and judge them on their merits.
It's a bad sign for HN that this facile dismissal has (as of this writing) twice as many points as this much deeper but also critical analysis:
Self interest aside, it's such a trollish statement that adds nothing to the real discussion, which is a) what's wrong with product x, and b) what my product y does to fix this. Such behaviour correlates reasonably strongly with incompetence, at least in my experience.
My problem with the grandparent post is more that the question was actually posed by a journalist, which is night and day different. They're sexing up the content to sell a story. So reframing the quote as being from the Clustrix guys is quite dishonest.
Had the quote actually been theirs, I'd have no problem laughing at their expense, although I realise such behaviour is not encouraged here.
That and cheaper hardware used more intelligently.
Even more sadly,I have found these topics, RDBMSes and hardware, to be areas of great ignorance and, perhaps therefore, fear among software folk.
More importantly, it will probably take a substantial effort to de-mystify these two.
Kinda reminds me of how Litespeed built a httpd.conf/.htaccess compatible apache replacement - it's drop-in ready but many times faster in comparison. But it's very pricey so you won't see Apache going away anytime soon (and people also switch to nginx for free if they don't need the compatibility factor).
Only PostgreSQL or a similar more featured database can doom MySQL, and I don't see this happening soon.
In fact, this product even strenghts MySQL in the market, by making it scale up for enterprise customers.
We were picking a vendor and had gone deep in talks with the MySQL folks to license their technology, and then Oracle stepped in and we picked somebody else.
Following the technology since then, it's clear that Oracle doesn't really know what to do with this thing. I think it would be a mistake right now for any company to pick MySQL for any major work without plans on moving it out of their product at some point in the future, it's simply on far too unstable ground as a database to rely on anymore.
Text obstructed; didn't read.
Just don't bet on these guys.
And, BTW, isn't this the Maureen O'Gara that wrote that "exposé" about Groklaw?
I wasn't going to buy anything Microsoft for the foreseeable future. Oracle got one more penalty on my book.
If they won't even provide proper service to hardware customers, I think they will be downright hostile to a competing bread and butter product that they now control.