Feb 26, 2010, 15:24 by dan_a
Very useful step through of development and scaling
Database optimisation
Remo Biagioni (Feb 26, 2010 at 11:00)
Talk at PHP UK Conference 2010 (English - UK)
A real life example getting more throughput with fewer queries.
Over the last year we've grown a database from a few hundred megabytes to just over one terabyte. The database is reported on and populated by a network of servers using PHP. As the database has grown we've had to look again our initial assumptions and ways of working. One table has over 2billion rows; 2.5 million rows every day are added to another table. This talk will cover how we use explain, foreign keys, normalising data without sacrificing performance, queuing and using memcache. And, how we've made the system run faster now than it did with a much smaller database.
Quicklink: http://joind.in/1453
Track(s): Sidetrack 1
Comments
Feb 26, 2010, 16:36 by Stereoket
Some very interesting points, but felt it was too much info to cover in short space of time. Would like to have seen more examples and time on them
Feb 26, 2010, 16:59 by Shrikeh
Started slow and SEO-y, then demonstrably got stronger and more relevant.
Feb 27, 2010, 19:45 by pavlakis
Some really useful points.
Scaling to 30+ servers? I wonder. How many of those where just VM's?
Still, pretty impressive.
Enjoyed the talk. Nicely balanced :)
Mar 1, 2010, 11:50 by Gabriel403
Enjoyed this talk very much, a great story type talk about how the company went from a horrible db to something smooth and quick
Mar 2, 2010, 00:13 by Letssurf
Talker spent too long talking about their company.
Talker didn't seem to full understanding what they had done technically a lot of the time. Mainly spoke about how big and great their servers where.
There was some useful information to take but not enough as the useful pieces where skipped over.
Feb 26, 2010, 12:40 by BinaryKitten
Talk had a lot of content I wouldn't have considered as part of the talk. Felt disengaged from the speaker and the talk subject. The main thing I picked up was "memcached go for it".