We have mostly finalized the Percona Live schedule at this point, and I thought I’d take a few minutes to introduce who’s going to be speaking and what they’ll cover. A brief explanation first: we’ve personally recruited the speakers, which is why it has been a slow process to finalize and get abstracts on the web. Sometimes you know someone’s a dynamite speaker and you discuss over the phone, and then it takes a long time to get a title and abstract from them. In many cases the better they are the busier they are, so this is expected.
Let me introduce just a few of the great speakers we have lined up for this event: Brendan Gregg, Dr. John Busch, and Vladimir Fedorkov.
Brendan Gregg is the crazy guy who likes to scream at a chassis full of disks and show graphs of the impact the vibration has on the disk latency. If you have not seen this, take a moment and watch it!
Brendan obviously knows his DTrace. He will be talking about how to really use DTrace with MySQL and InnoDB at a very deep level in the storage engine — not the high-level introductions you might have seen elsewhere. I do not think he will be yelling at any JBODs, but who knows. Either way, this will be a talk to attend.
John is CTO and President at Schooner, who makes a high-performance MySQL software appliance. (It used to be hardware-only, if you haven’t been keeping up with them lately.) His session will be about benchmarks of replication alternatives for MySQL. Hasn’t this been done before, and is it really worth doing again? Actually, no, I haven’t seen it done in a scientific way, and yes it’s an important session for two reasons. One, Dr. Busch is a very distinguished engineer who’s been building complex systems since before I was born. Two, this is a session that will teach you how to think about replication, not what to think about replication. Data replication is one of the hardest and most interesting problems in computing, and I think you will come away from this having learned a lot.
I also want to mention that this is a sponsored session. Our sponsors have been asked to keep it 100% technical, with no marketing, and I believe that every one of them understands the value in that request. Take a look at the speakers, titles, and abstracts, and I think you’ll agree with me: our sponsors are among the most accomplished engineers out there, and they have a lot of wisdom and experience to share. This session is a great example.
Vlad actually used to work with Percona, as a consultant, system administrator, and team lead. Now he’s the director of Professional Services at Sphinx. If I can paraphrase his session, I’d say it is about the things you don’t know about Sphinx’s capabilities. Many people think of Sphinx as a great search engine, but it’s much more powerful than just that. It has all kinds of features that let it do things MySQL is very bad at. For example, think about using MySQL to find the top N most important rows per group of rows, distributed across a cluster of servers, sorted in date order (not importance order) and with an indication of how many rows are in the whole set. That’s a nightmare of temp tables and filesorts and multiple queries and all sorts of things in MySQL — and that’s just for performing this task on a single server, not to mention a cluster of servers. Well, Sphinx is very good at this type of task! There’s a lot more, too.
This might be a good time to mention that Sphinx’s author Andrew Aksyonoff just published a book on Sphinx with O’Reilly! Congratulations Andrew. It’s in my (long) list of reading.
Well, that’s all for this time. If you are not signed up for Percona Live yet, I encourage you to get your tickets now. I don’t know quite what happened today, but registrations really jumped, and there is a good chance that we’ll sell every ticket we have. It happened in San Francisco and it could happen again here, even with the additional rooms we’ve rented (to expand to four tracks).
By the way, we are also planning an evening event with an open bar and substantial food, so you can stay around and mingle, without worrying about paying for drinks or rushing off to find food before you faint of hunger. We don’t have a contract signed yet, but we’ll keep you posted. This will be free for everyone to attend, even if you can’t come to the daytime event.