Oracle buys Innobase. MySQL between rock and hard place?

If I was MySQL, I would try to hire as many key Innobase OY engineers as soon as possible. Since the code is GPLed, the value is in the people who know it best.

However, this does put a crimp in MySQL's dual licensing model, because they won't necessarily be able to sell a MySQL/InnoDB solution commercially unless Oracle agrees.

"number of storage engines including MyISAM, Memory, Merge, Cluster and InnoDB. And with MySQL 5.0, we added the new Archive and Federated storage engines."

Jeez, how confusing. Plus there's the old ISAM. And Heap. And BDB. InnoDB seems to be the one to use among the older types, not sure how it fits in with the new ones. If Oracle tries to quash MySQL by drying up InnoDB dev, or extract lots of money via licensing, I think MySQL will just shift dev focus to another engine.

I don't really think Oracle and MySQL are direct competitors - they're at different ends of the market.

This can only lead to bad things for MySQL in the long run.

InnoDB was the only thing MySQL had going for it to seriously compete with Oracle, and Oracle knew it. Now that Oracle is acquiring InnoDB, in a year ot two, we'll see Larry Ellison but the major squeeze on MySQL AB to either pay up for licensing InnoDB or not include it at all.

Larry knows what he's doing. He saw a upcoming and serious competitor (i.e. threat) in InnoDB, bought it, and will squash it. Just like Microsoft has done with scores of other software companies. This will relegate MySQL back to running websites, non-transactional databases and data warehouses (which is OK, but not great).

A very sad day indeed.

The writing has been on the wall for a long time. In MySQL 5.1 MyISAM is supposed to have row level locking, I think Heikki saw the writing on the wall.

One of the amusing things about this is that Oracle is purchasing technology that they were vilifying earlier this century. I sincerely hope that Heikki squeezed them for every drop that they were worth.

Also, despite that InnoDB is one storage engine of many, it is still very important - it and MyISAM are the key database engines, while the others are (for most people) frosting on the cake.

Later on, Kaj makes an effort to *clam* the fears of MySQL users by saing that MySQL will continue to support all their users and work with Oracle as a "normal business partner."

clam?

Please recognize that your Chicken Little comments aren't helping anything or anyone. I'm certain that Oracle still intends to generate revenue as a part of the overall business model.

Licensing code is a big part of that - overpricing those bits (or refusing to license them) equates to no/lost revenue.

It's not as though there aren't a lot of bright people capable of innovating something different, better, and faster.

We just kicked Solid to the curb, in favor of MySQL, because of their [retarded] licensing policies and cost structure. If MySQL stops making sense, we'll choose another platform - and I'm 100% certain that Oracle will not be our only option.

I find threads like this very paradoxical. Doesn't Innobase have the right to sell itself (in fact, isn't that the basic operating premise of most start-ups)? Oracle overpaid for Innobase, though, I'm sure, long-term, it was a fair use of cash.

Oracle is akin to Cisco - the "partner" you hate, because they have your perverbial nuts in a vice, and keep raising their license and support prices every year. The "we'll tell you what you want and how much you'll pay us" mentality is so 1990.

I am not worried.

I guess the spellcheck didn't catch "clam." I'll fix that.

It normally takes a few years before you can poach employees from a private company in a merger/acquisition.

That said, what is everybody worried about? It's GPL. If Oracle does something unsavory, fork the code.

Jeffrey,

The problem isn't usage that fall under the GPL. The problem is non-GPL usage. You can't fork the code and permit people to use the modified version under any license other than GPL.

>> InnoDB was the only thing MySQL had going for it to seriously compete with Oracle, and Oracle knew it.

??? mysql does not compete with oracle, even in moderate datamart applications.

maybe oracle is trying to establish an alternate track product in the low-end market, perhaps even one day including mysql.

for the bulk of the big sales they do, mysql is likely not even in the bid process.

Ah hahahaha, poor MySQL :P

I might be wrong, but I actually think Oracle stands to benefit in helping MySQL grow (indirectly, by improving InnoDB).

Why? Because MySQL's market isn't the same as Oracle's: applications that would put MySQL in consideration likely wouldn't consider Oracle because of it's cost. So, who is MySQL really competing with? PostgreSQL, MS SQL Server, maybe even Sybase. If Oracle can help MySQL eat the lunch of those guys ... it means that Oracle can dominate the high-end DBMS market (the greater-than-$1M market) and MySQL can try to dominate the low-end DBMS market (the less-than-$1M market).

Smart play for Oracle and could be very positive for MySQL, too.

If you really think that customers weren't choosting MySQL when they'd otherwise have used Oracle (if MySQL didn't exist), you're kidding yourself. It happened quite a bit.

I've heard this from folks at Oracle, MySQL, and people who've contacted me for help and advice. MySQL was certainly taking customers from Oracle.

How many? I don't know. But it certainly didn't feel like a trivial number. Sure, they may not be Oracle best (meaning most expensive) customers, but a small customer today could be a big customer tomorrow.

MySQL's customers seem to be the small, start-uppy type companies with FOSS believers - livejournal, sixapart, and many others; and big companies like Yahoo; and a slew of hosting companies who offer MySQL to their customers. I've never heard of any banks, brokerages, credit cards, huge online stores (ebay, Amazon, etc.), etc. using MySQL.

I don't think Oracle believes now in Open Source technology. They just found the weakest point in MySQL and bought Inno Oy.

It's not just against MySQL, they also hit SAP, which works together with MySQL on porting MySQL to R/3 (and I don't believe that they use MyISAM or BDB) - and that might be much more important: A low cost DB in ERP market, that would be a nightmare for Larry.

The storage engine of MySQL is a good thing - I hope that MySQL will be able to acquire a new technology or to develop something on their own. For long term they should definetly drop InnoDB - community and customers don't believe in Oracle.

I thought the same thing, why didn't MySQL buy InnoDB years ago since it was such a valuable piece to their operation. I also can't believe that InnoDB would do the open-source community such an injustice. I know its business first, but as always with the open-source community they went against the grain with the business-first model.

I see it now, printed T-Shirts: "Oracle: Do The Right Thing"

If the worst happens and Oracle tries to squash InnoDB, there may already be such an alternative out there.

I wonder what it would take to add (and optimize) Postgres storage engine support to MySQL? I don't know exactly how current versions of MySQL and Postgres maesure up performance-wise, but PgSQL seems to have made steady progress on performance improvements.

Maybe this is a crazy idea, I don't know how technically or legally feasible it is, but I really like the idea of the two open-source communities uniting to battle Oracle.

To me it just looks like Oracle found MySQL's weakest spot and struck.

From a purely biz perspective it is genius - and it exposes a total failure in MySQL AB leadership to close this hole. How could they have let this happen to them? It's almost mind boggling and blunder that would have many boards of directors begging for a CEOs head.

>Interestingly, there's always been once piece of the InnoDB puzzle that's not
>available under the GPL: the InnoDB Hot Backup Tool. Without it, database
>administrators cannot backup their InnoDB tables without shutting down
>MySQL or at least locking out all transactions.

You can just use mysqldump with --single transaction, it places no locks and does everything in a consistent read

>I've always wondered why MySQL AB didn't buy Innobase Oy years ago.

I know they tried serveral times but MySQL doesn't have the kind of money Oracle has. Heikki tried to sell InnoDB years ago to Oracle before partnering with MySQL and they basically laughed at him. Now I bet they paid him a lot more.

That's all well and good for a small database, but really doesn't scale well at all. The relaod time on that sort of dump would kill you.

I use mysqldump for 2+ terabytes of data (though in tons os separate dump files). each dump is taken from a replication server and I insert the position attributed with the backup at the top of the dump file. I mainly use this because it is very flexible. I can create scripts that will restore and point in time recover any schema, move scema's between servers, etc.. With Hot Backup you can't move an individual schema to another installation which isn't flexible enough for my uses. Another advanage is you can stream the dump into gzip therby saving on disk IO and creates very small backups. To get around the reload time issue I keep replication servers up to date so they can become the new masters in case of a failure.

mysql could use firebird as storage/transactional/relational
engine in future ;) if oracle will cut InnoDB (who knows what will happen with it)
don't know about licences fit (mpl vs gpl)
and with it they could get oracle mode too (for free)

http://www.janus-software.com/fb_fyracle.html

I am the builder of Hotels.com and former chief architect, as well as ceo & chief architect for Aisus.com, a leader in insurance software. I can tell you without a doubt, MySQL and PostGre COMPETES with Oracle and Microsoft! Oracle does more, but you don't always need a thoroughbred, and in those situations, its nice to have a quarter horse in the stall that doesn't cost a fortune.

Casey Marshall said:
QUOTE:

If the worst happens and Oracle tries to squash InnoDB, there may already be such an alternative out there.

I wonder what it would take to add (and optimize) Postgres storage engine support to MySQL? I don't know exactly how current versions of MySQL and Postgres maesure up performance-wise, but PgSQL seems to have made steady progress on performance improvements.

ENDQUOTE:

It's an interesting idea, but I wonder whether or not time would be better spent converting older applications to just run on top of PostgreSQL natively, via an abstraction layer or something.

I have to agree that MySQL & PostGres compete with Oracle, but the reverse is not true. Oracle does not offer a low cost database solution. The real question here is, depending on what oracle does with InnoDB, will PostGres be able to fill the low cost, transaction safe hole that would be left if, MySQL loses the ability to distribute InnoDB. It has been my experience that PostGres is too slow for a medium volume application. MySQL is currently a perfect fit for a mid-sized transaction heavy database, and you don't need a dedicated DBA to keep it running- something that Oracle can't compete with.

Take a look at ANTS Software (ANTS) They claim to have a non locking data base that Will spin circles around any Oracle Configuration. Microsoft should take a hard look at them before Oracle goobles them up as well.

www.ants.com

Does anyone know if Innobase held any patents related to InnoDB? I would think Oracle could effectively stop any attempt to fork if they did.

QUOTE
It has been my experience that PostGres is too slow for a medium volume application. MySQL is currently a perfect fit for a mid-sized transaction heavy database
QUOTE

I'm now in the final stages of converting one of my projects from MySQL/InnoDB to PostgreSQL. The schemas are 99% the same, database sizes in MySQL/InnoDB version range from 5 to 25G. Pg's are normally 5-20% bigger (which is kinda expected as innodb uses index-based table storage vs pg's separate data/index files)

For last 2-3 weeks, I've been running the newer pg-based version on several production servers.

There can be a 0..200 transactions/second, average is 80-120 tps, 99.9% of which are simple index-based select/update/delete-s. The rest 0.01% are various long-running queries which do part- or full table scans (daily backups, stats analysis, data moves, etc).

Those 0.01% queries were one of main reasons for the switch. *any* long-running query makes all those index-based single-shot supposedly-ultra-fast queries go 2x-5x slower in MySQL/InnoDB. This was observed on various servers with 512mb..2g of ram (with proper memory configurations in /etc/my.cnf), 1x, 2x pentium4/xeon 2800+ CPUs, various IDE,SATA and SCSI drives, various Linux distributions.

The algorithm is simple: fire the backup, or a "DELETE" statemept (which can match/remove 100k+ rows), anything non-single-shot. Start doing "mysqladmin processlist" from the shell. Now, I can see 2, 3, 5, 10 active mysql connections executing those supposedly-"fast" queries (mostly coming from apache/mod_perl backend). I can even catch the session-loading SELECT-s (this is from a table which can have 100 rows max!). I'm 100% sure none of this is row locking related.

I knew PostgreSQL can do good, I've been working with it for last 2-3 years (vs 5-6 w/ MySQL/InnoDB). But the first results were simply amazing. In PostgreSQL, I can't even catch any of those "fast" queries (using "select * from pg_stat_activity ..."). None of "heavy" hdd read/write-intensive queries could make my "fast" queries go any slower. Web backend is as responsive as it was. VACUUM-s (via pg_autovacuum), backups, data imports (1-10m of rows), etc - PG is still 100% responsive.

Just my 2 cents..

Not really trying to turn this into a PostgreSQL vs MySQL forum- but I haven't had those issues with MySQL (except on xeon processors and limited memory). Also, I'm not sure that running mysqladmin to see open connections is a valid benchmark, maybe the problem is with mysqladmin or pg_stat_activity? I do know that my db is very responsive under the circumstances you are describing, but I'm running powerpc G5's.

Addendum-
Also, I'm not sure optimizing for 0.1% of your transactions is the right way to go

Just my 2 cents...

And all the talk about other storage engines (hooking up Postgres, Firebird, etc), does not help their existing customers/users who have all their data in InnoDB.

Nobody is going to trust their data to a 1.0 storage engine. Maybe a stagnate fork though (for a little while) -- though that doesn't help MySQL's commercial licenses.

And if a data migration is necessary, MySQL is going to lose customers/users when the users reevaluate available options. If Postgres is a viable option now for my company, do we choose to migrate to Postgres (a completely open source project) or MySQL and risk MySQL being bought or their storage engine provider being bought again (SleepyCat or whoever).

QUOTE
And all the talk about other storage engines (hooking up Postgres, Firebird, etc), does not help their existing customers/users who have all their data in InnoDB.

Nobody is going to trust their data to a 1.0 storage engine. Maybe a stagnate fork though (for a little while) -- though that doesn't help MySQL's commercial licenses.
QUOTE

You're 100% correct. As with any DB vendor issues, I think patience is going to be the way to go here. We'll have to see what Oracle is going to do. Personally, once 5.0 is the "Official" stable release, I'm going to buy a commercial license and hang on for the ride. At most I'll be out $500 bucks or so. If the Oracle deal goes sour, I'll have to re-evalute my options then.

I think they don't care about MySQL+Innodb's market. I think they do care about the buddy buddy relationship between MySQL and SAP.

Whither MaxDB? Probably.

I'm surprised that nobody has mentioned the SAP DB, which SAP "donated" to MySQL -- which rebranded it as MaxDB.

SAP acquired and developed MaxDB as a competitor to Oracle, because they were tired of lining Oracle's pockets with revenues from every SAP sale. SAP's database requirements are HUGE, and Oracle was the choice DB for a long time. Once Oracle entered into the Financials market, they became a direct competitor of SAP. MaxDB is an extremely scalable and robust database -- rivaling Oracle in many respects.

I should think that MySQL, which has not really embraced MaxDB heretofor, would now jump on enhancing and puffing up MaxDB with all due haste.

Just my $0.02.

Brett, agreed. Oracle has bigger fish to fry than MySQL (in this case anyway).

You're missing the big picture.

Think long term. Think google-scale clustering.

MySQL should focus on long-term clustering of their product line. Focus on performance on single system images up to clusters of several hundred machines.

They can do this with MySQL cluster and just add disk-based persistence.

salvatore Sorrentino said:

Addendum-
Also, I'm not sure optimizing for 0.1% of your transactions is the right way to go

ENDQUOTE

You might want to re-read that post. He wasn't optimizing for 0.1% of his queries. His point was that in MySQL, the 0.1% that took a long to run killed the performance of the other 99.9% that WERE usually fast.

The significance of Oracle’s acquisition goes deeper than what may appear.

Oracle just threw MySQL’s customers into a huge pile of uncertainty by buying the transaction engine that implements MySQL 5.0 advanced database features, which was announced only weeks ago in late September. What does MySQL have left without InnoBase? Their Classic edition is what, with a flat-file DB engine that certainly doesn’t support triggers or stored procedures or transactional capabilities.

These customers are also left to ponder what the community support around MySQL is as now clearly appears that all most critical development and everyday support is strictly contained within a few individuals. MySQL is more about business than community. And they make such a business blunder as to rely so heavily on external product that is now on the hands of Oracle. Now, what happens when Oracle decides to put a little squeeze on that license cost or heavily influence the evolution and support of that product? MySQL customers and prospects will have to reevaluate their database investments and analyze short and long term impact on their businesses.

People just need to recognize Oracle’s InnoBase acquisition for what it is, and MySQL customers need to have their wits about them over the next several months as this deal plays out. –Paola

> You can just use mysqldump with --single transaction, it places no locks and does everything in a consistent read

Wrong. It still does a flush tables up front, which must acquire a lock on every single table in every single database. Not a huge deal if you're mostly an OLTP database, really a huge deal if you're mostly an OLAP database with a steady transaction load as well.

MySQL was poised to become the Linux of database software in the next few years. That is one of the reasons Oracle bought Peoplesoft, etc. It will be a long time before there are any open source CRM vendors at the scale of Peoplesoft. I’ve been an Oracle customer for 15+ years. Based on that relationship, I have no reason to believe that Oracle has anything but their own interests in mind. MySQL users should be very worried.

I think MySQL needs to acquire SleepyCat immediately to counter attack to Oracle for revenge!!!

>Wrong. It still does a flush tables up front, which must acquire a lock on >every single table in every single database.

--single-transaction doesn't do this but --master-data does which is what your probably talking about.

By the way, I checked BerkeleyDB customer references from their web site. Their customer references are far more impressive than InnoDB's customer references. So why do we care about InnoDB at all? :-)

QUOTE

What does MySQL have left without InnoBase? Their Classic edition is what, with a flat-file DB engine that certainly doesn’t support triggers or stored procedures or transactional capabilities.

ENDQUOTE

Actually stored routines, triggers and views are not tied to InnoDB in any way. The main difference between InnoDB and MyISAM is transactional capabilities.

Facts: one of the most distinguishing Oracle features over its commercial competitors (MSSQL, DB2) is MVCC, which can benefit OLTPs enormously. The two large open source database engines with MVCCs are PostreSQL and InnoDB. SAP looking to be MySQL compliant (and SAP installs are very large OLTPs).

It's pretty safe to deduce that this will hinder the MySQL-SAP relationship if not bury it completely. SAP can now either look to try to expand MaxDB (which will take a while, coz hardly anyone has even heard of it) or try to do something with PostreSQL, which agains turns back the clock to zero.

I can see the biggest long term threats to Oracle as scalable databases which don't have to be as efficient as Oracle since they'd be pretty hard pressed to do that, but can easily beat its price/perf by using a lot more clustered hardware (and would still cost much less because of much lower licensing costs). However, there needs to be another component there - very strong support and integration, which is the usual deterrent to large, mostly conservative companies. I would put myself as not really a conservative but a pragmatist and I still dislike the idea of getting the main app from one vendor, a supporting tool from another, a GUI tool from a third vendor etc... And in the end most companies will look at the financial status of prospective vendors. In that sense open source adoption will be quite limited to certain departmental databases in certain companies. MySQL do have the potential to develop into a "trustable" vendor, but the InnoDB takeover is certainly a setback for that direction.

Going back to SAP, they ARE a large "trustable" company. In the large organisation I'm currently working at (a power company) they are even a "strategic partner". And they seem to have a lot of political leverage. And I could easily imagine this being the case in other companies as well. If one day MySQL/InnoDB really made it (say proper support, TPC-C benchmarks etc) it could easily take a lion's share out of Oracle's DB revenues.

Enough ranting for the time being...

Hardly anything galvanizes the Open Source community as a direct attack to technologies that it considers it's own. Oracle has just opened the door for MySQL to execute a marketing, recruiting and product building campaign, the intensity of which we have not seen in the community since the SCO Linux attack. How many would doubt MySQL’s ability to clean-room a Next-Gen engine in 12-18 months? In fact this is truly an opportunity to do something very different form the run-of-the-mill, “me too” technology that InnoDB was. And in the mean time – just to be funny – Microsoft’s SQL Server 2005 Express is free to redistribute and could be made into one of the MySQL federated storage engines 

SleepyCat rocks. I will totally ignore InnoBase from now on. Check out the Google case study of SleepyCat.

http://www.sleepycat.com/customers/pdfs/cs_google_1005D.pdf

Getting bit in the ass by cheap sh*t. Everyday I love hearing how crappy everything Microsoft is and how great everything not-Microsoft is. I use MySQL/Linux for my clients too cheap to buy Microsoft products. I tear my hair out everyday about how cheezy MySQL and Linux are.

Then my smart customers use me for Windows Servers, MS-SQL Server and C# programming. 10 years of Microsoft without a major problem. Not one crash! Fortune 10 clients, too! Critical healthcare applications! I build everything "from scratch" with Microsoft tools and never worry about 3rd party products. Security issues? Only if you are an amature! I demonstrated an ultra-secure system for the NSA and NASA on the Windows platform. Damn UNIX hippies about fell out of their chairs!

MySQL will never match SQL Server.

Thanks Man...

I needed a good laugh. BTW, it's amateur, which is also what we call people who build solely on Microsoft. (especially those who rely on it for stability or security)

Very interesting comments from all of guys. I enjoy comments more then this thread.Secondly I totally agree with Mustapha.

I been using Linux since 1998, and of course people who never used Linux will say that it is crap, but that's only in your brain, since it has been reduced to old-runny oatmeal.... :- ) by you only being on Windows. It's like owning and riding a Chrysler for the rest of your life, you are not aware of Honda, Toyota, VW or any of the other brand names, so of course you think Windows is the best!

HAHAHAHAHAHAHA

MySQL is great, we use it out of the box for a lot of our clients and those that need corporate like solutions, we point them towards Oracle. If Microsoft products worked on other platforms other than only Win-blows, then maybe people would use it more....

Just my 42 cents.

Ray