IBM Donates Java Database App. to Apache Foundation

the_pooh_experience writes "IBM has announced that it will open up Cloudscape by giving it to the Apache Software Foundation. Cloudscape, a small footprint Java database, is primarily used for small scale websites and point-of-sale systems. Its new, opensource name will be 'Derby.' Cloudscape (originally created by Informix, and purchased by IBM in 2001) has been valued by IBM at $85M."

"public domain" is not the same as "open source"

[samuel4242]

Geez is the NYT dumb. Putting something in the "public domain" means you relinguish control. It's owned equally by everyone. Choosing an "open source" license means you keep control. If you're careful about how you do it, you can even change the license terms a bit later. MySQL is constantly tweaking their terms because they're the sole copyright owner. Sure, it's available under the GPL, but they can tweak the terms for preferred customers. And they do! That's still their perogative because the code is NOT in the public domain.

Re:CloudScape

[Anonymous Coward]

Cloudscape (originally created by Informix, and purchased by IBM in 2001) has been valued by IBM at $85M.

Actually, Cloudscape was originally created by a
company called Cloudscape & the company was
bought over by Informix. Then Informix was bought
over by IBM.

Re:Tax Reduction?

[Grullie]

Cloudscape is hardly dead - it shows up prominently in Websphere Application Developer as the default embedded DB for EJB data. It feels a lot like MS Access - simple, quick, and dirty.

More techincal background

[TheToon]

The Cloudscape homepage: Cloudscape [ibm.com]
And more details with links to PDF documents: Features and Benefits [ibm.com]

I would guess that mysql would be faster for simple stuff, but Cloudscape could give it a run for it's money with support for more complex SQL.

Wouldn't know how it compares agains postgresql...

Re:503 : DELETE YOUR COOKIE

[Anonymous Coward]

delete your cookie and you can visit the site, it seems whoever touched the slashcode last, broke it somewhere in the login section (or the db is b0rken)
hmm no wonder its free

Re:"public domain" is not the same as "open source

[jusdisgi]

Nope...they won't put anything in without you signing the copyright over.

Re:MySql Competition?

[jaaron]

While I haven't used Cloudscape in a very long time, I imagine this is more competition to other Java open source databases like HSQL [sf.net], Axion, or McKoi.

Most of these databases are used by "embedding" them into an application (something not uncommon in Java programming), not as a standalone database server like Oracle or Postgres. Of course, like I said, it's been a long time since I looked at Cloudscape so it could have changed to be more of a standlone server.

I'm also surprised I haven't heard more about this in Apache, but I imagine it will first go through the Apache Incubator [apache.org] to sort out any legal issues and then end up somewhere in the Apache Database [apache.org] project. If anyone has more info, I'm interested to know.

Re:MySql Competition?

[Anonymous Coward]

I think that Slashdot runs MySQL. So saying that MySQL cannot run big webapps is a bit of an underestimation :-)

Anyway, personally, I see it more as a competitor to hsqldb, which is also an embedded java DBMS. Or sqlite, although the latter is written in C++. It has the potential to become popular as a DBMS embedded in applications, but I don't think it is usable as a real stand-alone DBMS, such as MySQL.

Re:Tax Reduction?

[jusdisgi]

That's precisely the way I read it...they don't see a need for this in their 'for money' software portfolio. Kept proprietary, there isn't really any way it would be profitable; this isn't something they can sell as a product...they might be able to embed it in hardware sales, but they can still do that. So, they just farm it out to the "free geek labor" and forget about it. It still has all the utility for them it did before, now they don't have to pay development costs, and furthermore they can write it off on both their balance sheet and their taxes. And the price they get to write off? Astronomical.

From the features page...

[Anonymous Coward]

...sounds much like SQLite to me!
But it's probably MUCH more mature. :) And slower. :(

Correction

[cdegroot]

Cloudscape was originally crated by Cloudscape, Inc. (I contracted for them at one time), which was later acquired by Informix.

At the time, it was a fairly complete and well-performing database with some nifty multi-database synchronization features, so even though I'm not involved in Java programming anymore this can turn out to be a quite interesting addition to Joe. A. Opensourcecoder's toolkit.

Re:Database written in Java?

[Tassach]

Java's reputation for slowness is based on a few factors, none of which is applicable here:

1. The GUI, partiuarly AWT, is horribly slow. This is unavoidable because there's no good (fast) way to do platform-independent graphics. However, since a database engine has no GUI components, this won't be a factor.
2. A lot of the perceived slowness of Java comes from the time it takes to load and initialize the JVM. This is a factor when launching a desktop app written in Java. However, since a db engine is a long-running (daemon) process, this is also not applicable in this case.
3. Early JVMs, particuarly before JIT (just-in-time compilation) took off, were indeed slow. Modern JVMs with JIT, once loaded, run (non-gui) Java code at a speed comperable to compiled C++. This is just 8 year old FUD that needs to die.
4. A database is almost never CPU bound. The very nature of databases causes them to be I/O bound -- I/O bandwidth (especially disk throughput) is almost always the limiting factor in database performance. Modern CPUs are so vastly over-powered for database work that even if it took twice as many cpu cycles to run the Java code as it would to run comperable C++ code, you'd still only be using a small fraction of the available CPU cycles. Excessive CPU consumption in a database is almost invaribly caused by user error -- poor schema design, missing indexes, poor SQL coding, etc.

Re:Database written in Java?

[LnxAddct]

You obviously haven't used java within the past 3 years or so.... Its speed on every platform I've developed on is no different then native speed, and in many scenarios its faster because of the many optimizations that java makes to your code and also from years of optimizing their own algorithms. This could be debated well... forever, but anymore if you need something faster then java then you should probably be using assembly. Speed and overall performance has only gotten better with the new 1.5 VM as well (It's now known as 5.0, and its still in beta but very useable).
Regards,
Steve

Re:Anybody used it?

[mauryisland]

I've been using Cloudscape 4.0 in a web environment for a couple of years now, with no database failures of any sort. Cloudscape has a good selection of utilities (bulk loader, CLI and GUI, etc.) It's picky about ANSI SQL, and it supports most of the SQL that I'm interested in, like nested queries, stored procedures, etc. I'm using it as an imbedded database (just presenting data, not writing anything while in production), so I can't speak to the speed in an OLTP environment, but for my purposes, I'm absolutely delighted with it.

What would be good....

[Phill Hugo]

If someone could develop and release an Open Source / Free Software Java Object Database that can house live java objects and provide searching and modification on these objects. Zope built atop Python provides just this for Python users and is used to good sucess on many web sites (Zope provides through the web editing of Python/Zope objects and has a very nice search interface for finding that object you require). Everything in Zope is a Python object and is stored in a fast, low resource object datbase - this includes all your web site and the methods to run on them. Its great, free and fast and is available from www.zope.org.

Re:Comparisons?

[bheerssen]

First off, Cloudscape runs on nearly every system in current use. MS SQL Express requires Windows (2000 or XP). It also requires a fairly beefy sytem for a light database. Check out the system requirements [microsoft.com]. Note that specialized MS platforms, such as MS Windows 2000 Datacenter Server (of all things) are not supported.

Finally, that MS crap is still beta software whereas Cloudscape is well tested, enterprise ready code.

Apache Corporation?

[gtrubetskoy]

Funny how the word Apache in the article is linked to the stock ticker for APA. (Or may be not so funny) For the record - The Apache Software Foundation is a registered non-for-profit 501 c3 corporation incorporated in Delaware, and as such it does not have stock but rather can hand out membership to make one a stakeholder.

Re:503 Service Unavailable

[LnxAddct]

Just remove your slashdot cookies and all is well again... go figure.
Regards,
Steve

P.S. Anyone know what the hell is going on over at Slashdot HQ?

Re:Tax Reduction?

[Anonymous Coward]

The Apache Software Foundation is a private operating foundation that is registered as a non-profit, charitable organization under Section 501(c)(3) of the U.S. Internal Revenue Code. This means that, for individuals within the U.S., donations to the ASF should be tax-deductible. [apache.org]

Re:MySql Competition?

[Frums]

Remember the "NDA Submission" discussion in the incubator recently which ended in a "handle it vase by case"? This is why. Target is the db project, and has been voted to be accepted there, so once it satisfies incubator, wheee =) -BrianM

Cloudscape was not written by Informix

[Anonymous Coward]

Cloudscape was bought by Informix. Which was later bought by IBM. IBM was after Infomix's customer base. Cloudscape was probably not interesting for them. Anyway it is a win-win senario for IBM and open source. IBM offloads a technology they probably do not want to maintain. In return they get a lot of goodwill publicity from it. Opensource gets a decent embeddable SQL database.

Re:Tax Reduction?

[Anonymous Coward]

It's hardly a dead technology. Cloudscape is to be used as the local data store in the next generation of IBMs messaging products (e.g. Workplace which is built on eclipse RCP - see www.lotus.com)

Lotus say that the Notes client will 'converge' with the Workplace client in the version 8 release timeframe so that'll put an eclipse runtime and cloudscape DB on most every corporate Notes desktop in the next 2-3 years.

What you're seeing is IBM seeding the developer marketplace with technology (Eclipse, Cloudscape) in order to reap dividends in the form of an established base of technologists familiar with the underpinnings of their commercial products.

You getting the picture?

Re:Hsql

[lokedhs]

We use Hsql, and it is indeed turly a good database. The greatest trength lies it its speed. It's incredibly fast as long as the database is small enough (so that it holds everything in ram).

The biggest problem we have had with it is that the only transaction level it supports in READ_UNCOMITTED, and that it doesn't support multiple parallel queries.

Re:What would be good....

[MemoryDragon]

Ok giving that I havent used OODBs in years. There used to be a OODB standard, the ODMG standard for java. Believe me the standard was disastrous. Missing standardization left and right, no conrol over the object scope, now handling of mass data on a sane level, the OQL was not really good for real world purposes etc... There were implementations of this standard, but you ended up always with code only geared to one OODB. Most java programs nowadays use OODB-RDB mappers, I assume Zope does the same. There are excellent ones, like Hibernate (which basically solves every problem the ODMG back than didnt) give one of these a try. Sun also comes out with something, done by former ODMG people, JDO, I dont know how this one stacks up to Hibernate, but I assume there arent too many mappers which come close to it or can surpass it. The third one I know is Torque from the Apache Projekt, mediocre at its best. The EJB also do this to a certain degree because part of their layer is direct object database mapping, but in my opinion they are total overkill.

Re:From the features page...

[cching]

Unfortunately (depending on your point of view), SQLite doesn't have java bindings (that I know of). If someone does know of java bindings for SQLite, I'd be interested!

Re:Foot in the door?

[Anonymous Coward]

Agreed... Informix rocked.

Even installation was an easy, no-brainer operation.

The documentation was also particularly excellent. IDS blew Oracle out of the water.

Cloudscape, not Informix, developed Cloudscape

[Anonymous Coward]

Cloudscape (the DB) was developed by Cloudscape (the company). Informix bought it, and IBM later bought Informix.

Re:Yawn.

[MORTAR_COMBAT!]

Another fast, small Java database is Hypersonic SQL [sourceforge.net].

hsqldb is a relational database engine written in Java, with a JDBC driver, supporting a rich subset of ANSI-92 SQL (BNF tree format). It offers a small (less than 160k), fast database engine which offers both in memory and disk based tables. Embedded and server modes are available. Additionally, it includes tools such as a minimal web server, in-memory query and management tools (can be run as applets) and a number of demonstration examples.

Another interesting, open source Java database is McKoi SQL Database [mckoi.com], a GPL-licensed Java database with all kinds of nifty features.

Mckoi SQL Database is an SQL (Structured Query Language) Database management system written for the JavaTM platform. Mckoi SQL Database is optimized to run as a client/server database server for multiple clients, however it can also be embedded in an application as a stand-alone database. It is highly multi-threaded and features an extendable object-oriented engine.

Things are getting interesting for JBoss [jboss.org] developers: JBoss ships with HSQL, supports McKoi nicely, and now we get Cloudscape thrown into the mix. Sweet.

Re:Anybody used it?

[Anonymous Coward]

We are a successful enterprise software ASP and used Cloudscape during our first 1.5 years before switching to a more robust database. Cloudscape powered our applications for 40+ customers globally. While it performed well given its small footprint, a huge problem for us was the fact that it does not release unused space -- the only way to do this is to run COMPRESS. With growing transaction volume we began running into random database balooning problems where the size of Cloudscape grew from 100MB up to 30 or 40GB depending on how long it was running. The fix required a COMPRESS command which takes anywhere from 1 to 48 hours depending on the size of the database and the amount of physical memory available.

Switching to Oracle and SQL Server eliminated this problem entirely. In addition, performance has been increased literally ten-fold just from the switch (no changes were made to our code or schema when this performance increase was measured).

While I would recommend Cloudscape for smaller, non-critical applications, it is not ready for real-world enterprise apps. I look forward to the improvements the open source community might bring -- from a cost perspective I'd certainly like to see us switch back someday.

Regards,
Matt

Re:Kinda similar to Berkeley DB for Java...

[Ragica]

Does anyone who voted the parent insightful have even the slightest clue as to what "SQL" means.

Wake me up when there's an SQL interface to DB.

Re:More techincal background

[bwt]

I was rather impressed with Cloudscapes features:
- store serializable java objects in table fields
- java stored procedures, functions, triggers
- hash joins in addition to nested loops
- subqueries and views
- transactions, four isolation levels
- row and table level locking (default = row)
- deadlock detection
- custom aggregators
- blob and clob data types
- cost based query plan optimizer
- multicolumn B-tree indexes
- query and lock built-in performance monitoring

To me, this feature set seems superior to MySQL.

Re:"public domain" is not the same as "open source

[Slime-dogg]

This type of program doesn't apply at all to the Linux kernel. The GPL is strange, in that instead of signing control of a copyright to a singular entity, you keep it... but what you wrote is "free." Free as in, you can't dictate to other people what to do with your copyrighted code, with exception of enforcement of the GPL. You've agreed to allow everyone to modify it, copy it, sell it, barf on it, etc.

The Kernel does have a group of controlling entities, most obviously Linus. Then comes all of the branch maintainers. Other than that, the "official" linux branches don't get modified without their approval.

The appearance of SCO would have affected Linux regardless of what license they are using. If IBM signed over the copyright to Linus, then SCO would still sue them. If someone decided to sign over copyrighted code written by me to someone else, I'd probably be upset too. Then again, SCO doesn't own any of the code that was placed into Linux, so the point is moot.

Re:Can I do that as well?

[aoteoroa]

1. Write any code
2. Put extreme high price on code
3. Give code away
4. Profit!!

Or am I missing something?

Yup. You're missing the fact that that IBM didn't write the code. Cloudscape wrote the code. Informix bought it, and then IBM bought Informix.

The $85 million price tag is not an abitrary figure but is the actual cost of the code when it was purchased in 1999.

Re:MySql Competition?

[kpharmer]

> MySQL is definitely ready for heavy loads

heavy transactional read loads for non-critical apps perhaps.

- Not heavy DSS/OLAP read loads though (where indexes don't work well and you want partitioning to bypass 95% of your rows). See Oracle, Informix & DB2 to see how this is done and the results it achieves.

- Haven't seen a proper benchmark but antecdotal evidence points to problems that MySQL has scaling to meet much write traffic. Postgresql, Firebird, etc on the inexpensive/free side appear to be better choices for this kinds of applications.

- Aren't online backups unavailable except through separately-licensed (and expensive) products?

- Then you've got the entire managability issue - on larger projects in which you desperately want the kind of functionality that MySQL AB has claimed that 95% of database applications don't need and which they've failed to support well: like database-enforced data quality constraints (referential, uniqueness, and check constraint declaratives). Add to that the lack of flexibility that comes from various missing features like views & stored procedures. Add to that the problems porting their non-standard SQL. Lastly, add to all of the above their massive list of exception-handling problems - in which errors silently fail.

Nah, MySQL is a nice little database. But unless 'heavy loads' means non-critical, read-only, index-oriented loads - I think that there are about a dozen better options available.

Oh yeah, and no - cloudscape isn't a competitor for mysql in general. They each bring different strengths to the table.

Re:Yawn.

[MORTAR_COMBAT!]

fyi, you can use MS Access as an ODBC database without actually installing (or owning) Microsoft Access.

1. Control Panel->Administrative Tools->Data Sources (ODBC)
2. Create a User- or System- level data source.
1. Add
2. Select "Microsoft Access Driver" and click Finish
3. Provide a name and use the 'Create' button to create an MDB file.
3. Connect to the ODBC database using your program of choise (Java, VB, TCL, whatever knows how to talk ODBC -- Java includes a JDBC/ODBC implementation) and create tables, enter data, etc.

This of course works only on Windows. If you're trying to provide a cross-platform tool with a DB, use something like HSQL (if you are programming in Java), or SQLite, Berkeley DB, etc.

Using PHP 5 as a development language, you would gain built-in access to SQLite and Berkeley DB anyway, I'm pretty sure Perl has all kinds of database stuff, and TCL definitely can talk to Berkeley DB (most of my toy cross-platform projects that need a DB are TCL/Tk and Berkeley DB).

Re:Database written in Java?

[MemoryDragon]

Swing starts to become slow as soon as it has to bypass the hardware acceleration functions entirely. Even using volatile images for double buffering in java2d can help immensely once the volatile images can be handled by the graphics adapter. You could see that in windows with the speed increase once you moved from jdk 1.3 to 1.4 it almost was five times faster. Alas the unix version was a little bit behind because using backbuffers and directly using the graphics cards ram is not that trivial in plain X, you have to bypass X or go the OpenGL route. Newere JDKs (1.4.2 rc1 from blackdown and 1.5 from Sun go that route)

Not really Informix

[PCM2]

I've never used Cloudscape, but coming from its Informix roots I trust this - to a certain extent, of course.

Fair enough, because after all its Informix roots only go so far. Informix didn't invent this product, it bought it. [javaworld.com] Strange that nobody has mentioned this before now.

Cloudscape has a relatively small market share among SQL databases, but it is popular in certain niches. It came bundled with Sun's reference implementation of J2EE at one time, too; I don't know if that's still the case.

Re:Anybody used it?

[Anonymous Coward]

What version of Cloudscape were you using? I've used Cloudscape on and off for prototyping various things and I've noticed that Cloudscape 4.0 uses space much better than previous versions. Also, I also suspect a 'storage leak' like that would definitely have been fixed since IBM began shipping it with Websphere and the Workplace suite.

It's cool to see there's a new version out for public consumption and contributing the source to Apache makes it even better.

Re:Pretty cool

[khuber]

(SimpleApp.java) has some of the worst formatting I've ever seen

It's just formatted with mixed spaces and tabs and your editor is probably converting the tabs to 8 spaces. XEmacs loads it okay if your tabs are set to 4 spaces, then you can select the text in the buffer and do a M-x untabify.

Otherwise, something like this should work

perl -ne 's/\t/ /g;print' SimpleApp.java > SimpleAppNew.java

There are supposed to be four spaces inbetween the / / .

503 Theories

[sbszine]

Anyone know what the hell is going on over at Slashdot HQ?

See this thread [slashdot.org] for wild speculation. Short version: there was a change to Slashcode that included making all of Slashdot available as a single RSS feed (rather than one feed per section as used to be the case), and on the hour RSS aggregators hit this single feed and somehow cause a 503 for logged in users.

Wild speculation, remember. Any actual solid info appreciated.

Re:Yawn.

[afidel]

Dear god don't use Access for anything. At the very least use JET, or preferably use the freely downloadable and redistributable MS Data Engine, it's basically SQL Server Lite.

Re:Kinda similar to Berkeley DB for Java...

[MyHair]

Wake me up when there's an SQL interface to DB.

Wake up! mySQL is a SQL interface on top of Berkeley DB [mysql.com].

However I agree the grandparent post is wrong in comparing BDB to Cloudscape.