RIT: High loads, my impressions.

A few days ago, the RIT conference was held in Moscow: High loads, which I wrote about recently when publishing the conference program. Mostly positive impressions: a two-day marathon with coffee and lunch breaks gathered a huge audience of gurus, specialists and students under a single roof. A wide variety of reports, which, contrary to their fears, were not just a retelling of advertising slogans and campaigns of speaker companies, is by no means!

This kind of conference is a great opportunity to talk among people who are committed to the same thing as you - you get real pleasure from talking with recognized masters, creators of the things that are used in hundreds and thousands of projects - these are nginx and sphinx, and much more another.

Unfortunately, the majority, but thank God not all speakers - these are ingenious technical experts, but inexperienced speakers - it was clear that not everyone can gently and fully convey to the audience the essence of the problem being raised. But to my pleasure I can note that the ability to speak in the speakers woke up - and they answered the listeners' questions with enthusiasm, apparently already having coped with the excitement and what is called, getting into the stream.

According to the reports, I will note with great pleasure Grigory Rubtsov from MySQL AB with the report “Approaches to optimizing the performance of MySQL” - just about a complicated, good report, excellent answers to questions that even poured on the sidelines as if from a cornucopia.

But for example, the guys from mail.ru about the Imagine framework - frankly did not like the company's own development, on which the company's servers work. Not because they didn’t speak well or didn’t prepare - simply by advertising their product, they do not give the opportunity to use it in their projects, that is, in fact, the whole report is written as a pitchfork - so I can advertise anything, and say - that this “anything” is the best. I don’t understand, why make a report in such conditions?

Report “Large-scale Jabber-cluster. IM and not only ”- frankly speaking, I personally did not like it - everything that was said there is accessible from the specifications, but it was said very, very much about some architectural features of the system used on some project (which they never said). Little.

The second day of the conference was marked in the morning by two interesting reports at once: this is Sphinx in the examples and tasks of Alexei Aksenov, so carried away that there was almost no time for questions - for which he paid - on the sidelines he was simply torn to pieces :) Report “Organization of asynchronous task processing »Yakov Sirotkin from Yandex made me wonder if we are doing everything right. Well done!

The next noteworthy Yandex Monitoring Report by Maxim Lapan. Yandex monitoring system is a rewritten zabbix, moreover, Maxim said several times that zabbix was basically redone in the interface, and talked about functional alterations several times :)
By the way, in this system, after some kind of failure, Maxim (or administrators) receives an SMS -message so i think afterYesterday , I had to turn off the phone briefly - there were a lot of SMS :)

In general, everything was interesting :) I would like to hear your opinion about the conference.