Yep. Oracle is scared of pretty much any other database (properly not sqlite, but otherwise).
They are so scared that they prevent anyone with an Oracle license from publishing a benchmark of it. Which can only mean that they suck at everything except playing golf.
To be honest it makes some sense in their case. Just look at the amount of blog posts that compare how fast databases are by running an insert of 1M rows, then a couple of selects and declaring a "winner" based on that... I'm sure Oracle wouldn't like any "tests" like that published.
I don't think that is really their concern. Even if Oracle comes out on top in benchmarks they lose. "We're spending how much to get a x% performance increase?" Once numbers are attached a rational decision can be made come buying decision time. Oracle wants/needs that decision to be based on qualitative arguments, "our space unicorn is better than they're blue elephant."
I think you're confused. All of the major DB vendors have clauses saying that you can't publish benchmarks without their written consent. Not surprisingly that supports a hearty confirmation bias. What's the confusion?
>They are so scared that they prevent anyone with an Oracle license from publishing a benchmark of it
I believe every single major database vendor does this. It's called a DeWitt clause. The arguable justification is that it avoids misleading or incompetent benchmarks from slandering their product, but in reality it's so the only benchmarks published are the ones that favour them.
I recently was looking to add Microsoft's new AppFabric (sort of like Redis) to an install, but strangely was finding zero performance metrics on it. I finally plowed ahead to discover that one of the first EULA demands is that you don't publish any benchmarks. I cancelled the install and moved on.
Products that can't bear comparison instantly enter the land of the dubious.
I feel your pain. I've decided that after my current big Java project, I'm going to move into HTML5+CSS3/Ruby web apps. I've seen what people are doing with JS these days, and I wonder why I'm still writing my UI in C++/Qt and Java. I find Sunoracle's lackadaisical treatment of Java 7 and JavaFX unacceptable.
Agreed. I've been enjoying dabbling lately with some of the new JVM languages but with Oracle at the helm I'm really not sure if this is a good investment of time.
I think the open-source community really undervalues Postgres. MySQL does what it does well enough but the Postgres team has done a phenomenal job building a robust, fully-featured open RDBMS. Hopefully the replication features of the 9.x release will get it into more server rooms.
Man, Oracle works /hard/ to maintain their bad reputation. I mean, they do some good work, look at BTRFS, but it looks like their general public relations to nerds is really, really bad.
To some extent, you have to make a choice to focus your marketing on managers or to focus your marketing on Engineers... but even so, it sometimes seems that oracle goes out of it's way to piss off the Engineers.
I think it was more about Solaris support than just the mere fact of having another build server.
Sun Microsystems - and for a short time its new owner Oracle - had provided three member servers to ensure PostgreSQL was stable on the Solaris operating system.
With the end of OpenSolaris I think it would be wise for the open-source community to write off Solaris support. Time spent supporting Solaris is time better spent making Linux a better server OS and making sure that things like Postgres work as efficiently and reliably as possible on Linux.
Linux will never get all enterprise features of Solaris. That OS, broken into a myriad of distributions, driven by lammers who think they are making smth great while they're not - is it gonna become your OS of choice for a mission critical system?
Where did you see "the end of OpenSolaris"? Is all that FUD made by Linux users who I so happy to blame the competitor? See http://www.oracle.com/technetwork/server-storage/opensolaris...
Could it be that this was an accident? I can imagine someone at Oracle going "Sun guys, what do these servers do?". The Sun guys, some of whom may have already left Oracle, some of whom may have shifted departments recently, don't know and say: "they don't seem to be used for anything, according to the administration. There's some weird postgres stuff going on there, but it doesn't seem to do anything useful". "Well OK", the Oracle guy says, "then we can use those machines elsewhere".
I don't think so. It's not uncommon to run Oracle on Solaris + SUN hardware. Not a large market but one with high profit margins. Why would Oracle want to kill a cash cow?
That is true, but Oracle recently has been pushing Linux, not Solaris. The Exadata, which is Oracle's first attempt to to build a database appliance on Sun hardware uses Linux, not Solaris, for example.
"It's very dangerous to let anyone fly under you. If you have the cheapest, easiest product, you'll own the low end. And if you don't, you're in the crosshairs of whoever does."
"In technology, the low end always eats the high end. It's easier to make an inexpensive product more powerful than to make a powerful product cheaper. So the products that start as cheap, simple options tend to gradually grow more powerful till, like water rising in a room, they squash the "high-end" products against the ceiling."
I think Apple is the exception. They ate the other mp3 players and they ate the other mobile companies' profits. So it seems like if the high end is cheap enough, it can eat downwards.
I'm a big fan of Postgres and have a hard time overcoming my visceral dislike for Oracle. But I suppose this makes perfect sense for them. Businesses protect themselves from real and potential threats. If I were an Oracle shareholder, this is probably what I would want and expect them to do.
Except it does nothing to protect Oracle. The cost is so inconsequential somebody is going to provide hosting. All they achieved was a couple days (hours?) of hassle for the pgsql people and pissing off some devs. Zero competitive advantage was achieved while confirming to developers they can't be trusted. If it was an accident, Oracle employs humans. If this was purposeful, Oracle employs idiots.
I will. I run PostgreSQL on Solaris at the day job and it runs very well. (Most things run fine on Solaris so long as you avoid X.) We use a lot of Sun hardware where I contract, and it will be very annoying if I end up having to port my software to MySQL to make Oracle happy.
42 comments
[ 4.1 ms ] story [ 86.4 ms ] threadThey are so scared that they prevent anyone with an Oracle license from publishing a benchmark of it. Which can only mean that they suck at everything except playing golf.
Then what's this? http://www.tpc.org/tpcc/results/tpcc_perf_results.asp
I think you're confused. All of the major DB vendors have clauses saying that you can't publish benchmarks without their written consent. Not surprisingly that supports a hearty confirmation bias. What's the confusion?
I believe every single major database vendor does this. It's called a DeWitt clause. The arguable justification is that it avoids misleading or incompetent benchmarks from slandering their product, but in reality it's so the only benchmarks published are the ones that favour them.
I recently was looking to add Microsoft's new AppFabric (sort of like Redis) to an install, but strangely was finding zero performance metrics on it. I finally plowed ahead to discover that one of the first EULA demands is that you don't publish any benchmarks. I cancelled the install and moved on.
Products that can't bear comparison instantly enter the land of the dubious.
http://harmony.apache.org/
I think the open-source community really undervalues Postgres. MySQL does what it does well enough but the Postgres team has done a phenomenal job building a robust, fully-featured open RDBMS. Hopefully the replication features of the 9.x release will get it into more server rooms.
To some extent, you have to make a choice to focus your marketing on managers or to focus your marketing on Engineers... but even so, it sometimes seems that oracle goes out of it's way to piss off the Engineers.
The have them anyway, and they build PostgreSQL anyway when it's uploaded to debian (although obviously that's not every commit).
But with ccache a rebuild is not very expensive.
Sun Microsystems - and for a short time its new owner Oracle - had provided three member servers to ensure PostgreSQL was stable on the Solaris operating system.
Several Oracls DBAs I know used to say "if you want to run Oracle you might as well run it on Solaris"
"It's very dangerous to let anyone fly under you. If you have the cheapest, easiest product, you'll own the low end. And if you don't, you're in the crosshairs of whoever does."
http://www.paulgraham.com/start.html
"In technology, the low end always eats the high end. It's easier to make an inexpensive product more powerful than to make a powerful product cheaper. So the products that start as cheap, simple options tend to gradually grow more powerful till, like water rising in a room, they squash the "high-end" products against the ceiling."