Oracle's MySQL buy a 'fiasco' says Dovecot man Mikko Linnanmäki A co-founder of the widely-used IMAP server Dovecot has outlined his three rules for open source success, in terms Larry Ellison may not enjoy. “The first rule is don't sell your company to Oracle if you want to keep your product alive,” he told World Hosting Day in Singapore yesterday. “The second rule is also don't sell …
Sun hard 3 assets that were of value to Oracle -
MySQL
JAVA
Sun Hardware
If Sun did not own MySQL, you can bet your bottom dollar what Oracle would not have touched them with a barge pole. I think the Gent's real gripe is the development and release cycle has taken somewhat of a nose dive and they are not playing nice with others, fact is MySQL is no longer innovative, instead they are playing catch up with forks such as MariaDB.
A few months ago we made the decision to dump MySQL in favour of MariaDB, probably the best move we've in a while.
Oracle is institutionally incapable of innovation
Maybe nowadays, but they were the first to market with a relational database and came up with some pretty cool features for it over the next twenty or so years. Plenty of products have surpassed Oracle DB now though, not least in terms of usability.
This post has been deleted by its author
> Are you really suggesting that Oracle thought up Relational Databases ... or even SQL first?
No, of course not, but it brought them to minicomputers and vastly improved them. Like I said, IBM invented them, then sat back and admired them.That's why Oracle is somewhat more widely used than DB2.
What official histories? The ones authored by Larry?
There were some interesting articles in the IEEE "Annals of the history of Computing" in late 2013 and eraly 2014. They looked at the DB industry as a whole, and had articles on Oracle (written by an early Oracle employee) and IBM by an IBM employee. They make interesting reading.
I think the Gent's real gripe is the development and release cycle has taken somewhat of a nose dive and they are not playing nice with others, fact is MySQL is no longer innovative, instead they are playing catch up with forks such as MariaDB.
I wonder why he was even asked about the deal. Does Dovecot use MySQL in a large way?
I disagree that MySQL isn't faring well under Oracle. Then again I would disagree that MySQL itself was ever innovative. It seems to me that Oracle is actually working quite hard at making MySQL a reliable product by squashing bugs, while admittedly introducing new ones, improving the tools and making it something businesses will be prepared to pay for support for. Sun never really had a plan to monetise it. Yes, Oracle isn't playing nice with some of the forks but they're anything like as unreliable and unpredictable as MySQL was in the past then I'm not really surprised.
Where Oracle did do a disservice to itself was with Hudson and OpenOffice.org. But the world recovered and moved on.
Charlie Clark: " It seems to me that Oracle is actually working quite hard at making MySQL a reliable product by squashing bugs"
...yet the biggest improvement in uptime,memory use and speed in my MediaPortal setup was the result of swapping in MariaDB. Working hard perhaps, not achieving much for the effort.
I found this open source taliban always funny. Google is always good, its proprietary API of course are better than an RFC standards like IMAP, Google never does evil... (of course Exchange APIs are very, very evil, instead).
Meanwhile Oracle is the "Great Satan", and of course the many $$$$$$$$$$$$$$$$$$ MySQL developers got in the sale (to Sun...) were not important in their decision.
Sure, nobody would ever buy Dovecot, but I wish to see someone offerinh him not an Instagram or a Whatsapp, but some hundred of million dollars anyway, and see how he reacts....
Oracle or others, I frankly find hypocrite this "don't sell" unti they offer you millions (or billions) of dollars, and then all your "principles" go away. Nobody forced Widenius & C. to sell MySQL.
But I understand what is the new business open source business plan now:
1) Make an open source project appealing
2) Sell it for $$$$$$$$$$$$$ to X company, which usually people think is bad
3) Fork it
4) Tell people to use your fork and no the original product because X is baaaaad, very baaaaad
4 bis) If you can, cry X needs to give you back the project rights, because X is baaaaad, and doesn't deserve them because it's ruining the holiness of the project (this may not work, though)
5) Get more $$$$$$$ from the fork, if your plan works, while still enjoying those got at 2)....
I found this very hypocrite.
If meanwhile you plaud Google for introducing its own proprietary API and protocols to access mail instead of standard, "open source" ones, well, you're really like a "taliban" - you act on irrational "religious belief" only.
Proprietary API normally means that the owner won't let anyone else know what they need to know in order to call the code, or to write a compatible library or server with the same API but a different codebase. For example, Microsoft's obfuscated calls in the WIN32 API which only MS Office was able to use (to make sure Office couldn't run under WINE), or Microsoft's very expensive (and eventually failed) battle with the EU to keep the fine details of its AD fileserver protocols secret and thereby keep Samba/CIFs from becoming a full competitor).
If Google has published the documentation of its new API and isn't trying to prevent other folks interfacing to it, it's not proprietary. It's just an alternative to IMAP that might become tomorrows standard, if it attacts sufficient interest. (Just as IMAP was once a fancy new alternative to POP).
No, proprietary APIs means you own them and design and implement (and change) them as you like.
'Proprietary' never meant 'reserved' or 'unpublished' API. And who warrants you Google published all the APIs and there are none reserved for Google only use? Google itself?
IMAP was designed by IETF without being designed for a single company interests. Of course Google wants you to use its own protocol instead of standard ones, so you get locked in and it's easier to make send your data to Google for its own interested use. It's exactly what MS tried...
But again we see that MS is evil, Oracle is evil, Google is good...
BTW: CIFS was an old MS name for its SMB protocol. SMB predates AD, and is not tied to it.
If meanwhile you plaud Google for introducing its own proprietary API and protocols to access mail
"Accessing mail" is just another way of implementing a distributed filesystem that has the form of a holdover from the era of servers that were like Raspies at the far end of a serial line.
It's time to let go and make a mail client a gloss on top of Coda for example.
Open Software often sounds like communism to me. Nobody owns it. Therefore, nobody gets paid for it. Therefore it's not of any (real) value to anyone at all. It's communism in software.
It would be funny if someone offered to buy *his* business for a few million. It's funny how communists suddenly turn into capitalists when a big fucking cheque is waved in their face!
"Hi Bill, don't you have some philantrophic project to manage, you old scumbag?"
Bill's given up the philanthropy game after GQ awarded Tony Blair the 'philanthropist of the year' award.
I think Bill's waiting for the current Middle East "peace" to end so he can try and repeat Tony's success there next.
The reason nobody will buy Dovecot is because it's a pretty basic tool that anyone with a decent development team could write on their own in a very short time. The Dovecot guys have built a solid IMAP server, for sure, but at the end of the day it's just an IMAP server.
"The Dovecot guys have built a solid IMAP server, for sure, but at the end of the day it's just an IMAP server."
There are plenty of IMAP server implementations available. Most of them are crap and more than a few of them will corrupt mailboxes.
Dovecot is good because it's well written and was structured from the outset.
There's nothing stopping it being dual licensed just like mysql was.
A more pressing issue is what happens on mail clients when users insist on having 60Gb/10 million items of mail in various folders (this is a real world scenario).
Not likely. None of of the experts want to talk about it, but many companies (mis)use their mail systems like that. At the moment I don't think the outfit I'm working for is quite that bad (25G limit on the GMail cloud), but I've worked at them before and provided service to even more.
Must be a bad day in Reg-world. Is any of this very controversial? Oracle bought Sun because they had some cool technology (e.g. Java), a ton of really top-notch talent and sucked giant fried donkey balls at making a profit. By the way they also vacuumed up an annoying little upstart called MySQL. OK, nothing too surprising there.
Anyone who knew of Oracle's biz practices would have predicted that for all their protestations, the open-source assets would be monetized in some way. As it happened, Oracle took a rather ham-fisted approach to some of them (eg the CI tools) and people forked. Whether you like Oracle or not, it is their right to do as they please with the assets...just as it is a saving glory of the open-source licensing schemes that we, the community, can get around it safely. The end.
The legions of mysql users that weren't paying Sun/MySQL for support still aren't paying Oracle for MySQL support? Oracle didn't care about those people anyway, probably in excess of 90% of them were not going to consider paying Oracle for support regardless.
You're right that people looking for a cheap/free database backend for their blog weren't going to splurge kilobucks per year on Oracle. These people represent new entrants into the DB customer base.
There are however the 10% you refer to, old Oracle customers who've started to use MySQL, Postgres or nosql in situations where Oracle's features and support aren't required. Every MySQL installation there is a direct loss to Oracle. When making the decision between 1) paying Oracle 2) paying for "free" DB support or 3) self support the choice is less and less the first one. It's in Oracle's interest to steer MySQL development away from "enterprisey" features to shore up their market share, or at least sell those features at a premium.
While I'm mostly inclined to agree there's is a problem. Whereas 2 years back "the cloud" pretty much guaranteed it was outsourced, there are apparently now "private clouds" which might or might not be at the other end of a distant fiber connection. So if it is a private cloud on your premises you might be OK. Or not, depending on your backup solution, disaster recovery plan, and a dozen other things.
All in all, where as once I thought I understood what people were talking about, I'm now a bit cloudy on the whole thing.
So very true. By the end of the day, it's not like their data is stored in some fine mist with a cable dangling from it. The data has to be on an array somewhere, so they end up paying for the purchase, support, connection and on top of that, overhead and margin of their provider.
If you use a considerable portion of your storage, shifting to cloud makes no sense whatsoever. If somebody offers you a lower price per byte stored and transferred, they're cutting corners somewhere. The problem is, you don't know where. And finally, if they grossly underestimated costs or overestimated profits, their business model falls apart, they fold and you're left without a provider and usually without any way with which you can recover your data.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
