c0t0d0s0.org

Don MacAskill (the CEO of SmugMug - yeah CEO and geekdom isn´t mutually exclusive) did an interesting experiment - he used a Opensolaris system for one of his database replicas. And his experiences were really positive (besides of the usual pet peeve we already working on ):

I’m a Linux geek, have been since 1993 (Slackware!). All of SmugMug’s datacenters (and our EC2 images) are built on Linux. But the current state of filesystems on Linux is awful, and it’s been awful for at least 8 years. As a result, we’ve put our first OpenSolaris box into production at SmugMug and I’ve been pleasantly surprised with the performance

The configuration has a really interesting speciality. He used compression.

Lo and behold, it worked! We’re getting a 2.12X compression ratio on our DB, and performance is keeping up just fine. I ran some quick performance tests on large linear reads/writes and we were measuring 45.6MB/s sustained uncompression and 39MB/s sustained compression on a single-threaded app on an Opteron CPU.

I planed to write an article about this "Sun will close the source of Mysql" nonsense. Alec Muffet summarizes the real story of this Enterprise/Community Edition discussion on his blog in Whats the SQL? by linking to an authoritative source: A Slashdot article written by Marten Mickos, the former CEO of Mysql AB .... wait ... a CEO writing at Slashdot ????

I found two really interesting interesting presentation. At first Brent Paulson wrote an really interesting presentation about practical Opensolaris Security. It´s a good overview about the possibilities of Solaris to increase the level of security at your installation. Ben Rockwood of cuddletech.com wrote an good preso about the usage of dtrace in conjunction with mysql.

Kishore Singh Thakur published with "MySQL InnoDB Performance Tuning for the Solaris 10 OS" an excellent article about tuning mysql and Solaris. A must-read.

Wired reports about the Mysql AB in Sun Pays Up the Wazoo -- $1 Billion -- for MySQL. The article itself is not so interesting, it´s one of this terms and conditions articles, which were legion in the last few days. Nevertheless there is a comment in the article , which explains my antipathy in regard of the most surplus profession - analysts

"Less than 1 percent of MySQL users pay," wrote Global Equities Research analyst Trip Chowdhry. "Users are in the habit of getting free stuff and it's impossible to break the habit. . . We don't see Sun (being) able to monetize it."

They don´t see an opportunity even a blind man at midnight would recognize. At first: The tens of thousands of users, who run their blog on a mysql database on their debian server are commercially uninteresting, there are more interesting to build an opportunity. But there are users, who run their websites or their applications on mysql and need a supported database, but don´t want to spend a fortune for oracle. And these people want to pay. And at the size of the Mysql community 1 percent of users are a large group of customers.. It´s simple, these "research analys" simply don´t understand the model of open source, i doubt they understand the IT business as whole.

Even when we are note able to monetize the Mysql at large, but with Mysql we have a powerful enabler to talk with companies about our products, about our servers. We get important even for a pure Linux on HP, Dell and IBM shop, as we own the M in LAMP now. And this is an important point to monetize on the 1bn $.

Paul Murphy comments in Fun with Sun in Forbes the latest articles at Forbes.com about the Mysql AB deal. This article contained so much errors, that i´m decided i won´t link it. But Paul summarizes it quite good:

In summary pretty much everything Greenberg says here is filtered through a soul deep mask of ignorance - but our problem is that a lot of our bosses are in his audience - and I don’t expect Forbes to be publishing corrections any time soon.

Jonathan answers some questions about the deal in his newest blog entry: in a vortex

The second-last paragraph matches my experiences as well:

And having listened to about 10 customers face to face over the past couple days, I've heard only one comment, made consistently - "Congratulations, this is absolutely fantastic news for all of us!"

I´ve read many articles and comments in varius internet publications and communities which speculated about the future of Mysql now Mysql AB will be acquired by Sun. There are some concerns about "Will Mysql be an Solaris-only thing", what will be the focus of the further mysql development. I thought a little bit about this fears. At first ... this thoughts are my own opinion. I don´t now much about this deal. The deal surprised me as well as most people out there. I don´t think that i have much more informations like the ones circulating in the net. Okay ... let´s start ....

I don´t know, what Mr. Dvorak has done in the past to justifiy, that such underinformed and underthought articles like this are still published by someone. This is the same guy, who tried to told us that the iPhone battery has only a reach of 40 minutes and years before he wrote, that the computer mouse will be a gadget nobody will ever use, that Apple will switch to Windows ... and other pearls of computer journalism.

Jason from Digitar describes in his blog how he is backing up his mysql databases with ZFS. He published his backup script was well.

Jonathan, do us only one favour in conjunction with the Mysql AB acquisition: Don´t allow marketing to rebrand Mysql as Sun Java System Relational Database Management System - Web Edition. Pleeeeease !!!!

TPM wrote a really positive comment about the acquisition of Mysql AB by Sun:

This is arguably one of the smartest moves Sun has made in a decade, and probably rivals the decision it made in the mid-1990s to buy the carcasses from Thinking Machines, Kendall Square, and the Cray Sparc-based server business.

He compares the deal with the most important deal for Sun so far, the acquisition of Cray Business Systems Division. Cray developed a 64 processor system, the Cray CS6400, at that time based on the SuperSPARC architecture. SGI bought Cray in 1996 but didn´t want an additional processor architecture in their portfolio. Thus the Business Systems Division was sold to Sun. The successor of the CS6400 was based on the UltraSPARC architecture and had the internal code name Starfire. The official name of this system was Sun Enterprise 10000. This system was really important to build up our reputation in the datacenter.

The deal with Mysql AB has similar potential. At least when you combine the ubiquitousness of Mysql and Suns reputation into considereration. You won´t get an die-hard mainframe user to mysql, but more people will think twice before using a large wad of cash for buying a proprietary database.

Today, Jonathan announced, that Sun will acquire Mysql AB. This is an incredible move: At first, this helps us against the ridiclious pricing model of Oracle in the advent of systems with 256 cores. And as i headlined this entry: Sun is now the M in LAMP. This is nice position. You can read more about this deal in the press release

PS: An interesting side effect ... does this mean, that Kris will become a Sunnie in the next few months ?

Update: Kris already wrote about it in his blog (sorry, in german only).

Although a little bit advertising, the blog entry of Jignesh Shah points to an interesting fact: You can easily spend hundred times the server price for the database license. The business case to use a proprietary database should be a really good one. There are still companies out there, who use databases like DB2 or Oracle for all data because they did it so for the last 10 years. Opensource RDBMS like Postgres or Mysql may have not all functions of an Oracle, but you should ask yourself: Is this advantage really big enough to justify 40.000$ or 15.000$ per core? In my humble opinion this price tag justifies a dual database strategy.

John David Duncan of Mysql wrote an article about mysql on ZFS. His conclusion:

ZFS introduces remarkable ease and flexibility of administration, without any real cost in performance. At its worst, in these tests, ZFS performed almost as well as UFS with Direct I/O. With InnoDB, the ZFS performance curve suggests a new strategy of "set the buffer pool size low, and let ZFS handle the data buffering."