Previous month:
March 2006
Next month:
May 2006

PBXT Engine


At the MySQL Users Conference I met with Paul McCullough, the developer behind the new PBXT transactional engine.  Paul is a super sharp developer and also a very nice guy and he obliged me by posing in a photo as part of my "storage engine stalker series" of photos including Ken Jacobs of Oracle, Jim Starkey author of Falcon, and Paulo Lubet from Solid. 

It will be interesting to see in the coming months how PBXT and other engines flourish in the community.  We'll do our best to provide more engines, samples, documentation and ideas on the MySQL Forge as well as the developer zone and forums.

Solid Storage Engine


Our partners Solid Information Technology have joined our Certified Engine program and will be making available an open source engine that plugs directly into MySQL.  Solid brings robust OLTP database expertise and their system has been proven in telecommunications and finance for many years.

Solid is showcasing their new engine for MySQL at the MySQL Users conference booth 510 and will also have a presentation at the conference on the technology.   The picture above was from our offices in Cupertino (coincidentally MySQL and Solid are in the same building) when Paola Lubet, VP Marketing from Solid, joined us to celebrate the availability of their new engine.

More information about our certified engine program will be available at . The beauty of the MySQL pluggable storage engine architecture is the availability of many third party engines from commercial vendors and the open source community.

Congratulations, Jonathan!


Jonathan Schwartz has been promoted to CEO at Sun Microsystems as co-founder Scott McNealy leaves the CEO post after 22 years to become chairman.  McNealy's colorful style sometimes put him at odds with Microsoft, Intel and other industry giants, and some accuse him of not moving fast enough to adapt Sun's strategy and workforce to the challenges of the post Internet boom of the 21st centry.  Rumors had been floating about McNealy's departure recently and despite negative comments around Sun's recent financial losses, overall this is a good move for Sun. 

In the last few years, Schwartz has become considerable force inside of Sun, leading the charge of Sun's software strategy and their push to open source.  His promotion to COO was seen as a likely step towards taking over the reigns from McNealy at some point.  Now with the return of retired CFO Mike Lehman, the executive team will be in a strong position to make the changes necessary to return Sun back to a position of strength.

Sun has made a series of acquisitions in both software and hardware in the last couple of years and has launched a very successful line of high-performance Opteron-based servers.  Their move to open source Solaris, which would have been unthinkable a few years ago, has shown Sun is not afraid of making bold moves.  They've got the hot boxes and good open source roots, now they just have to execute on the strategy Schwartz and McNealy put in place. 

Codename Falcon


At our user's conference this week, Jim Starkey will be discussing his latest database engine, codenamed Falcon.  Falcon brings to MySQL Jim's 30 years of data management experience to create an engine that is optimized for modern scale-out applications with lots of memory. 

Falcon is ACID compliant with a multi-generational architecture that guarantees server-enforced referential integrity.  It's based on MVCC (which Jim essentially invented) so it keeps lock contention near zero, ensuring very high performance.  Falcon uses advanced B-Tree indexing and includes crash recovery.  Jim's approach has also been to make sure the engine is easy to use and automatically manages storage allocation so there's no need to manually reorganize things.

We'll be presenting Falcon in the context of the overall server roadmap Tuesday late afternoon at the MySQL Users Conference and then Jim's session is Thursday afternoon.

GarbageScout (updated)


When we talk about the future of the web, we often think of web sites from the big guys, like Google, Yahoo or huge phenomena like the FaceBook, or NeoPets.  But open source is actually enabling a new generation of entrepreneurs and individual developers to create applications that are in some cases just garbage.  And I mean that in the best way possible.  Let me explain.

A few weeks ago when I was at LinuxWorld, Bill Hilf from Microsoft demonstrated a very cool site called GarbageScout.  If you haven't seen it, you really need to check it out; it's the perfect example of a Web 2.0 "mashup" that combines Google maps information with photos that people post of random garbage that has been abandoned on New York streets.  When I lived in New York 20 years ago, I was always amazed at what you would see people throwing out: furniture, art, musical instruments, antiques, various junk.  Now you can view it all in your browser.  GarbageScout also now shows the latest junk in San Francisco and Philadelphia too.

James Nachlin, the developer of GarbageScout built the site to encourage repurposing and recycling. After all, one man's junk is another man's treasure. More interestingly, GarbageScout is possible because of the repurposing of Google Maps blended with a MySQL database, some PHP code for the user interface, Perl scripts and JavaScript. James is a professional sys admin and he uses MySQL to communicate information through the email parser, the geolocator (that translates text addresses into latitude and longitude) and the web front end.

Without Google Maps or an open source stack, GarbageScout would probably still be just a dream. Ok, maybe not a dream, but you know what I mean. And in turn GarbageScout is licensed under the Creative Commons license.


I exchanged some further email with James Nachlin. The GarbageScout site runs on Red Hat Enterprise Linux 3, using a Linux 2.4 kernel along with MySQL 4.0. Originally James wrote most of the system in Perl (his favorite language) but later re-wrote some pieces in Java.

His primary motivation for building the site was his self-admitted fascination with garbage and the fact that he's cheap. He likes things that have a history and age to them and he likes the archeological aspects of garbage.

Software Development Marathon


I think the marathon is a good metaphor for software development. Not in the sense of an enduring hardship (though that's often true), but in the sense that both represent very tough, but ultimately achievable goals. The analogy is not perfect though; in particular running a marathon is a solitary effort whereas software development usually is not.

I've run ten marathons in as many years, including Boston Marathon this past Easter monday.  My first marathon was ten years ago, when I was working at Borland on Delphi.  That was a tough project, but ultimately very successful.  Delphi went on to generate around half a billion dollars in revenue and is still in use today, which is kind of remarkable. 

  • Write it down. It's amazing to me how galvinizing it can be to set a goal in writing.  That's the first step.  The opposite is also true.  If you don't set goals, you won't achieve success.
  • Follow a plan. You can no more show up at the starting line in Boston and expect to finish than you can work on the next big software application without an understanding of what you're building.  Make a spec, a project plan, an architectural design and measure and monitor your progress.
  • Make progress.  If you make progress every single day towards your goals, you will get there.  There will be days where little progress is made, but it is better than no progress.  Want to play the piano?  Stop talking about it and start playing. 
  • Care. It's easier to make progress towards goals that you care about personally.  Don't live someone else's dream.  Don't just go through the motions.  Pick goals that are personally important.
  • Believe.  This is perhaps the toughest thing of all.  Every software project I've been a part of and every marathon I've run, has had parts where you just want to give up.  Someone quits.  The code sucks. Your legs are dying. Your family resents you, whatever.  Sometimes the only logical thing to do is quit.  But the people who are ultimately successful defy that logic time and again.  They have the tenacity and belief in themselves that keeps them going even beyond hope.  If you do not believe you will succeed you will not.  That's why athletes mentally rehears the events they compete in, to develop that faith that gives them strength when it is tempting to just give up. 

Ok, this was not the usual open source discussion, but I hope it is interesting to a few people.  For those interested, you can view some low-res photos I took from my Treo 650 while running the Boston marathon on Monday.  The photo at the top is the Citgo sign, a famous Boston Marathon landmark which is first visible at mile 23, a full two miles on the horizon. There's a final uphill climb to get to the sign at mile 25 and even when you get there, there's still another 1.2 miles to go. That may not seem like much, but I can tell you, you need to believe in your goal to keep running at that point.

Jim's New Storage Engine


Jim Starkey, who joined MySQL in January as part of our acquisition of his company Netfrastructure, will be speaking at the MySQL Users Conference on April 27 in Santa Clara on the new transactional storage engine that he's developed.  Jim is a veteran of the database industry for more than 30 years.  He pretty much invented some of the most significant innovations in the industry including blobs, event alerters, multi-version concurrency control and more.  He's also an avid boater, pilot and all 'round nice guy despite his reputation for being a "big bad wolf" in the open source world.   

We'll have sessions on many storage engines including MyISAM, InnoDB and others.  Last year we had 1300 attendees at the conference and many tutorials sold out.  We're on track to surpass last year's attendance and the exhibits area is virtually sold out.  If you haven't registered yet, I encourage you to do so.  Register in advance and you save $100 over the walk-in price.

Soul of a New Machine


I was talking with a friend of mine recently about how different areas have tried to rival Silicon Valley for influence in high tech.  But the valley's leadership wasn't always so.  In fact, if you go back twenty-five years, the "Route 128" area around Boston was the hotbed for much of the technology innovation around mini-computers with companies like Apollo, DEC, Data General, Wang and others.  And that in turn led to the development of many of the innovations in the early microcomputer industry with companies like VisiCalc, Lotus Development Corporation, Spinnaker Software, Javelin and many others.

Tracy Kidder told the story in his Pulitzer award-winning book Soul of a New Machine published in 1981, describing the development of a new 32-bit platform that would compete head-to-head with the DEC VAX.  Instead of covering the story from the top, by interviewing CEO Ed de Castro, Kidder told the story from the perspective of the engineers who made it happen.  He describes how Tom West, the head engineer, led a skunkworks project that would eventually become the Data General Eclipse/MV minicomputer. 

Kidder captured the spirit of dedication that it took to build an launch a computer system better and with more drama than any other book that has followed.  Data General was around for 30 years with many breakthrough products and was eventually acquired by EMC in 1999 with revenues of more than a billion dollars annually.  If you haven't read this book, you need to. It's a classic.

LinuxWorld Boston Panel


LinuxWorld Expo in Boston last week featured a number of better than average keynotes and panel sessions.  The best one I saw was one called "Is Linux Killing the Enterprise Software Model?"  The panel was moderated by Larry Augustin and featured open source CEOs Marten Mickos (MySQL), Marc Fleury (JBoss), John Roberts (SugarCRM) and Peter Levine(XenSource).

Overall the panel was good.  Marc Fleury was his usual lively self, slamming IBM and old school enterprise software companies.  Marten Mickos was the elder statesman of the panel, providing a good perspective on balancing the needs of enterprise customers and the open source community.   Other good points made:

  • Open source software must be easy to consume
  • Consumers of open source need to try it out themselves to see if it works
  • It's not just about "free" as in "no cost" you must have access to source code
  • Open source is both a development approach and distribution model
  • There is an impedence mismatch between the needs of IT (stability) and the Community (Release early, release often)
  • VCs are willing to invest in open source because the sales model is cheaper than traditional enterprise software companies
  • Sooner or later, ever sector will be impacted

Otherwise, LinuxWorld Boston was smaller and less exciting than last year.  There seemed to be fewer attendees at the session than in past years.  There was lots of interest in virtualization, but no really huge announcements.  Notably absent from the tradeshow were HP, IBM and Sun, which was presumably still reeling from their latest layoff.   

MySQL Will Support SAP


Shai Agassi, president of SAP's products and technology group said yesterday that he expects MySQL to be certified to run SAP applications by the end of the year.  This is an excellent endorsement not only of MySQL, but of open source technology in general.  SAP R/3 is the benchmark when it comes to the most sophisticated and demanding enterprise application and the testing process is exhaustive (if not exhausting!)  And even for users who may never need to run SAP, it's good to know that MySQL will be able to standup to all the testing.

It's something we've been working on for quite some time with SAP and it's nice to see that the top execs are aware and supportive of our efforts.  Agassi has sometimes been misunderstood in his comments on open source, but in our view, they've always been a huge supporter.