Sleepycat Software, makers of Berkeley DB, have announced the availability of Berkeley DB Java Edition, a pure Java database based on a write-once, append-only log architecture. Berkeley DB Java Edition is embeddable and built for speed and high concurrency while preserving transactional data integrity.

Berkeley DB Java Edition features full ACID transactions and recovery, record-level locking, schema neutrality for data storage in its native format, and zero administration. It offers the same storage services as the popular Berkeley DB engine, but the new product was completely redesigned in Java to take advantage of Java's portability and services such as deeply integrated threading and NIO. Berkeley DB Java Edition includes full source code for easier integration and debugging.

Checkout Berkeley DB Java Edition and read the press release.
Read Java Wakes Up Sleepycat on internet.com.

If they do it it is worth a look.

I wonder if Subversion can use this so their Eclipse plugin need not be compiled to use a specific version of the Berkley DB DLL

The eclipse plugin is a client to the SubVersion Engine. Sure, the Subversion engine could be modified to use the Java edition using JNI (OR a C binding to the java berkeleyDB).

What you're probably trying to say is " Is there a way to use a JAVA binding to the Subversion engine ?" Since the Subversion engine is continuously changing, writing a Java interface module to interface with the changing DLL versions would mean that this Java Interface Module would also have to keep changing ??

BR,
~A

Ported Blitz JavaSpaces from the native version to the Java version in under a day. Had a test release out for several months with my users and so far they're reporting no problems.

Good mix of features, good performance and saves you the pain of native libraries (a big thing for my non *NIX users).

Well worth checking out......

For those people worried about the so-called "viral nature" of the GPL, take a look at this. From the product licensing page (my italics):

The open source license for Berkeley DB permits you to use the software at no charge under the condition that if you use Berkeley DB in an application you redistribute, the complete source code for your application must be available and freely redistributable under reasonable conditions. If you do not want to release the source code for your application, you may purchase a license from Sleepycat Software. For pricing information, or if you have further questions on licensing, please contact us.

However, it sounds like some sort of alternative license is available, but they don't publish the terms.

That kind of a lameo license wont hold in court anyway. They should just make up their mind about making money on it or having a public source. These viral licenses are ridiculous.

That kind of a lameo license wont hold in court anyway.

Great! It's always good to get these opinions from qualified lawyers. Thanks for clearing that up.

(sarcasm tags not included)

It needs pointing out. That type of license is very annoying and very frustrating. There are cases where the source simply cannot be disclosed. (Because of other agreements and contracts). It's very egotistic of code authors to slap this kind of license to whatever they are releasing. The only thing it leads to is people either stealing the source and ignoring the license, or not using it. People use a specific tool/library because of its merits, not because of the particular taste/religion/gender of its author(s).

I happen to like Death metal, I'm going to invent a Death-license. Anyone using works I produce have to listen to death metal and watch pr0n out loud in their offices everytime they read the source. Anything else is to be considered a violation.

Yeah, stfu.

Then pay for a license or don't use the software. Those are the terms underwhich the software is available to you. Don't like it, tough. If you want to issue your stuff under your "Death-license" that's your choice but you can't force others to do what you want with their code.

Then pay for a license or don't use the software. Those are the terms underwhich the software is available to you. Don't like it, tough. If you want to issue your stuff under your "Death-license" that's your choice but you can't force others to do what you want with their code.

So... authors releasing their code as GPL can force others to jump through fire hoops with their code, but I can't? I like my Death-license. Makes me very productive and I wish for all of you to be as Zen as I am in my quest for the ultimate source code blahblah.. are you sure you don't see the similarites?

It needs pointing out. That type of license is very annoying and very frustrating. There are cases where the source simply cannot be disclosed. (Because of other agreements and contracts). It's very egotistic of code authors to slap this kind of license to whatever they are releasing. The only thing it leads to is people either stealing the source and ignoring the license, or not using it. People use a specific tool/library because of its merits, not because of the particular taste/religion/gender of its author(s).I happen to like Death metal, I'm going to invent a Death-license. Anyone using works I produce have to listen to death metal and watch pr0n out loud in their offices everytime they read the source. Anything else is to be considered a violation.Yeah, stfu.

There is a major flaw in you thinking. People do use it. People are developing with it. People do license it. The things you mention, un-ethical people will do. But they will do that no matter what or who they are working with. The Berkeley DB stuff is good. Use or don't.

Robert

It needs pointing out.

This topic has been debated at excruiating lengths in other many locations by people much more intelligent than yourself. I think you're probably missing a fundamental point somewhere, but I can't be bothered trying to explain anything to you.

For those people worried about the so-called "viral nature" of the GPL, take a look at this.......However, it sounds like some sort of alternative license is available, but they don't publish the terms.

That's not quite correct. There's a clause concerning "redistribution" and what that means which determines whether or not you may use it for free. Basically, if you're using inside your enterprise for building your own systems there's a good chance it won't cost you......

That's not quite correct. There's a clause concerning "redistribution" and what that means which determines whether or not you may use it for free. Basically, if you're using inside your enterprise for building your own systems there's a good chance it won't cost you......

Dan, you are correct. However, if you so much as give your code to a contractor, or a parent company (something which most subsidiaries would be obliged to do), then you must make the source code for your entire application available to the world. I don't consider that a "good chance", as both of these situations would apply to my organisation.

If you want it GPL you play by the GPL rules. If you need a commercial license, contact Sleepycat. It is your choice. If you have situations where you are not sure, contact Sleepycat, as they are a good developer friendly company.

I think it's a really great open-source model that works for their little niche. I believe their motto is "if you make money, we make money".

I know some of the developers and some of the corporate culture. These guys are into doing things very, very right (they spell check thier code comments, for god's sake). They think in terms of terabytes of data, and hundreds (if not thousands) of concurrent threads. Really!

I believe in using the best tools available, proprietary or open source. If I was in a situation that demanded a ultra-high speed transaction cache, I'd be going right to Berkley DB and paying them their due. Nothing home brew is going to be as fast or as reliable.

These guys are into doing things very, very right (they spell check their code comments, for god's sake).

That is music to my ears. I admit that I haven't looked at the product, nor have I seen the code. But I do believe that every project should follow firm engineering discipline with clear coding guidelines, frequent code reviews, and even spell checking of comments. So it is great that Burkley DB team believes into doing things very, very right :-)

Best of luck,
Dmitriy.

If you want it GPL you play by the GPL rules. If you need a commercial license, contact Sleepycat. It is your choice. If you have situations where you are not sure, contact Sleepycat, as they are a good developer friendly company.

Hi Robert,

I wasn't aware that it was available under the GPL. It appears to me to be their own license, which is much more restrictive than other GPL'ed databases such as mysql, postgres or firebird (which as far as I'm aware don't require you to redistribute your application's source code if you merely use the product in your redistributable application).

If you want it GPL you play by the GPL rules. If you need a commercial license, contact Sleepycat. It is your choice. If you have situations where you are not sure, contact Sleepycat, as they are a good developer friendly company.

Hi Robert,I wasn't aware that it was available under the GPL. It appears to me to be their own license, which is much more restrictive than other GPL'ed databases such as mysql, postgres or firebird (which as far as I'm aware don't require you to redistribute your application's source code if you merely use the product in your redistributable application).

Now I see your point. I read up on the license. It is not GPL but is OSI approved. So I guess it still carries my point. You still have the choice of license and I would still encourage you to ask them about any particular situations you have.

Robert

The open source license for Berkeley DB permits you to use the software at no charge under the condition that if you use Berkeley DB in an application you redistribute, the complete source code for your application must be available and freely redistributable under reasonable conditions. If you do not want to release the source code for your application, you may purchase a license from Sleepycat Software. For pricing information, or if you have further questions on licensing, please contact us.

if you so much as give your code to a contractor, or a parent company (something which most subsidiaries would be obliged to do), then you must make the source code for your entire application available to the world.

Just because you are required to make the source code "available and freely redistributable" to a parent company or contractor, does not mean that they will necessarily choose to redistribute it.

There is no chain-letter clause in the GPL or similar licenses that says something to the effect of:


/*if you can read this comment, you must send the complete source code of your application to 10 of your best friends or something bad will happen to you.*/

This would have been really useful if it included a JDBC driver. For my purposes it's not practical to lock myself into a proprietary data access API, so I will have to continue using HSQLDB.

A JDBC driver doesn't make sense, because its an embedded database, not a relational/SQL database. It's a very, very fast, very, very scalable, transactional byte store.

It's very useful for building local caches; there are a number of very prominent web sites that make use of Berkely DB ... some of these sites get millions of hits per day, and need to do 40 or more queries just to display the home page, yet still get ultra fast response times.

Isn't it a good fit for a JDO API? What would be the challenges of putting JDO API on JE?

A JDBC driver doesn't make sense, because its an embedded database,

How can it be used in J2EE? J2ee is a distributed system, there are many JVMs and physical machines. How can an embedded database be used concurrently by different JVMs?

A JDBC driver doesn't make sense, because its an embedded database,

How can it be used in J2EE? J2ee is a distributed system, there are many JVMs and physical machines. How can an embedded database be used concurrently by different JVMs?

Implement some trivial RPC wrapper like session bean and use it in the same way as you use any component. If you need then it must be possible to implement resource adapter for "global" transactions too.

A JDBC driver doesn't make sense, because its an embedded database,

How can it be used in J2EE? J2ee is a distributed system, there are many JVMs and physical machines. How can an embedded database be used concurrently by different JVMs?

Implement some trivial RPC wrapper like session bean and use it in the same way as you use any component. If you need then it must be possible to implement resource adapter for "global" transactions too.

So, Sleepycat would be "data store", the wrapper would be "database". So you implement "home grow" database. The cat is too sleepy.

Session bean is not a good example. Only *one* stateful session bean can be binded to sleepycat. What is going to happend when the stateful session bean is fail-overed?

JDBC would be useful.

Acelet

What is going to happend when the stateful session bean is fail-overed?

It is not a good idea to store "enterprise" data in B+Tree this way,
I know some people use it for "temprory" data to detect duplicates or for log file processing. Cache for content is a good use case in web applications (it will be faster than OS file system with many small files).
RDBMS has more good use cases than plain index file, this file can not replace database with SQL frontend and "enterprise" features. BDB is good (I know classic version), but you do not need to use it for everything, it is good to solve very specific problems.

I wonder how Berkeley DB compares to PERST in terms of speed & features.

I've been experimenting with PERST (http://www.garret.ru/~knizhnik/perst.html) a few weeks ago and I've really been amazed by its speed (with a DB containing 1 million of multi-kilobytes records with multiple indices).

Regards,
Sebastien.

http://www.jroller.com/page/spetrucci

I didn't look too hard, but I didn't see any pricing for the non-GPL version on Sleepycat's site. I did find this however:

"Berkeley DB Java Edition is available either through an open source or commercial license, with commercial license fees starting at $40,000 for a single-user, proprietary application."

http://www.infoworld.com/article/04/06/16/HNsleepy_1.html

"Berkeley DB Java Edition is available either through an open source or commercial license, with commercial license fees starting at $40,000 for a single-user, proprietary application."

Gee, a $40,000 per user license. What a bargain.

"Berkeley DB Java Edition is available either through an open source or commercial license, with commercial license fees starting at $40,000 for a single-user, proprietary application."

Gee, a $40,000 per user license. What a bargain.

Uh, the license states that it's $40,000 per single user application you create and distribute without source, not $40,000 per copy you distribute of your application. Similar in price to a Visual Basic for Applications license.

is it Relational or OODB ?

Neither, it is a B-tree database: Write-once, very-fast-read (updates effectively supercede older records rather than altering them).

You're permitted structured records, but this isn't relational. Many OODBs are based on this sort of back-end, but this is not, itself an OODB.

Does it have sql like language for searching ?
The main reason relational databases are used is that programmers do not have to code search functionality. Otherwise programs become too complex.
How would you do something like this simple query ?

select distinct a.name
  from authors a, books b
 where a.city = 'London'
   and b.title like 'H%'

Damn, no way to edit your message.
Here's correct sql

select distinct a.name
  from authors a, books b, authors_books ab
 where a.author_id = ab.author_id
   and b.book_id = ab.book_id
   and a.city = 'London'
   and b.title like 'H%'

Unlikely. Usually you access this kind of database either using a key (like in LDAP) or through sequential reads. If you need a relational store that permits queries like your example, you would probably not choose this one.