Xaprb

I am ridiculously far behind on reading blog posts since the MySQL conference. I like to lag somewhat behind, because then when I read the posts I get to see some comments. But I’m much farther behind now than I want to be. This is mostly the fault of the MySQL replication, Performance Schema, NDB Cluster, and InnoDB developers, who just announced a large number of improvements in the upcoming MySQL 5.6 release. How long did it take them to draft all those posts? I hope they keep blogging steadily between releases, too. In the meantime, I’m too far behind. It’s a good problem to have.

I split my time last week between the IOUG’s Collaborate conference in Orlando, Florida and O’Reilly’s MySQL Conference & Expo in California. The contrast was stark. For me as a MySQLer, Collaborate was a dud. On the other hand, the MySQL conference O’Reilly puts on is superb. It is vital to MySQL as a project and as a community, and it follows that it’s vital to MySQL’s business success. Oracle needs to participate to make it a success in the future.

MySQL at Collaborate had good speakers and content, but no one there is interested in MySQL. MySQL is just from a different world — it is a curiosity at an Oracle conference. Also, as a speaker, sponsor, and attendee, Collaborate was a giant frustration. I can’t recommend it to anyone. (These comments do not reflect on the work that MySQL community members did in recruiting and organizing the MySQL content at the Collaborate conference.) In particular, the experience of submitting talks was so disrespectful of my time and efforts that I have no desire to repeat it. I will spare you the many details, but I’m not alone. Many people I know echoed my sentiments and said they are not willing to present there again.

On the other coast, a lot of people were planning to fly east to Collaborate midweek, but upon arriving to California, I overheard many of them discussing the distasteful experiences MySQLers in Orlando were having, and canceling their eastbound flights to stay in California. This was wise.

The O’Reilly conference was much the same as in past years, with two notable differences. One, there was a lot of content about other open-source databases. This was good, although I wish that the whole conference could have grown to include them, rather than shrinking MySQL to make room for others. Two, the expo hall was downsized from previous years, a very worrisome sign indeed. I think it’s obvious to everyone who’s been around a few years that the lack of sponsorship from MySQL themselves is probably the biggest factor in play here. In years past, MySQL made a large contribution to the conference not only as a sponsor, but in publicizing the event and getting exhibitors, sponsors, and attendees there.

I can’t read Oracle’s mind and I have no insider knowledge to share, but from what I understand, Oracle’s well-intentioned strategy was to have the MySQL conference-goers find a new home in tracks annexed to events such as Collaborate and Kaleidoscope, where Oracle would be interacting with an organization they were familiar with. Now that we’ve seen this in action a few times, it’s crystal clear that it won’t work. The open-source mindset is stifled in those events, and I predict that self-respecting speakers won’t submit any proposals in the future. As a result, the business environment that’s built up around MySQL will dry up and blow away if the strategy doesn’t change. That’s why I think it is mandatory for Oracle to come to the table with O’Reilly and make their MySQL conference continue to happen. I would not be surprised if the people making the decisions at Oracle are simply unaware of the unique atmosphere at the MySQL conference, and how important it has been for MySQL in the past. Without this event, I think Oracle’s investments in MySQL could go to waste.

On the bright side, I think that if Oracle puts their shoulders to the MySQL conference’s wagon wheel, it could not only return to its previous vigor, but grow far beyond that. And if Oracle wants it to be focused on MySQL instead of including all sorts of other databases, I would not mind that at all. Any talk whose title made it clear that it wasn’t about MySQL was rather poorly attended, so although I thought it was nice to have a diversity of databases there, in truth it wasn’t all that different from jamming MySQL into Collaborate.

Update: Someone I respect indicated that the discussion in the comments was causing more trouble than it was solving, so I am sorry, but I think it’s best that we honor this request. I’ve disabled comments on this post.

Those of you who’ve been following my recent work on modeling system scalability might be interested in this. (It’s not my work, by the way. I’m just trying to ski in the wake of Neil Gunther.)

I’ve measured quite a few systems that have some strange bubbles in the scalability curve. As I explained in my talk on Thursday, systems don’t always follow the model precisely, because of their internal architecture. Systems sometimes behave differently at specific points because, say, an internal structure gets filled up and allocates a larger array to hold more of whatever it is, or CPU scheduling changes to balance threads across cores differently, or any of a number of other possibilities that are sometimes hard to understand or predict. Yes, this is hand-waving. But although it could sometimes take a lot of work to explain these kinds of things, it’s easy to observe and measure them in action, so the phenomenon is clearly real. VoltDB, for example, does not follow the scalability curve at 1 and 2 nodes in the cluster because 1 and 2 nodes are magic numbers. In computer science we usually say that there are three types of numbers: 0, 1, and many. Turns out it’s a little different for VoltDB.

So, for those of you who are curious about this, I now stop blathering and simply direct you to Neil Gunther’s blog to read more.