| You Will Be Amazed | Are you a gamer? | Free Cash - Click Here | Like the social link networks? | |||
| How To Take £100 And Turn it Into £20,000 - So Easy - Look | Don't miss our expert selection of awesome gaming news | Discover The Easy Secret To Making £1000's This Month! |
Then you'll love our social news site. The best of digg, reddit.. | |||
| www.roulettesystemreport.com | www.eplayr.com | www.roulettesystemreport.com | www.deliggit.com |
Comments on How YouTube scales MySQL for its large databases
This is a good one for those who take an interest in admin, advocacy, audio, database, dba, Google, lecture, linux, memcached, mysql, performance, python, replication, scalability, scale, scaling, sql, youtube
-
- How YouTube scales MySQL for its large databases
- In mid 2006, YouTube served approximately 100 million videos in a single day. To maintain a website of that scale, one would imagine YouTube has hundreds of DBAs. But in fact, there are just three people that make it all work.
Like most research labs, we rely on MySQL whenever we need a database. And like most (I’m guessing, here), it’s common to overhear something like the following in our lab — “We really need to replace MySQL with Oracle or DB2 in X so it can han…
-
[43 del.icio.us posts][47 reddit points]
- [Preview comments
]
In mid 2006, YouTube served approximately 100 million videos in a single day. To maintain a website of that scale, one would imagine YouTube has hundreds of DBAs. But in fact, there are just three people that make it all work.“According to him, the three important reasons for YouTubes scalability are Python, Memcache and MySQL replication, the last having the most impact.”scaleHow YouTube scales MySQL for its large databases By Tim Finin on Friday, December 28th, 2007 at 10:04 am. Like most research labs, we rely on MySQL whenever we need a database. And like most (Im guessing, here), its common to overhear something likeBlogspam.As much as a I LOVE PostgreSQL over MySQL, the criticism I received was, “If the database crashes, who are we going to call? MySQL is a company. PostgreSQL isn’t.”Try postgres + slony, you’ll save a lot of hair. But keep in mind that normal replication doesn’t scale as well as what they call ‘log shipping’ in the slony world.If you read the documentation for slony it says it’s not suitable for use over WANs and it’s not multi master.Slony does look excellent - we’re not looking to scale massively so it should handle things very comfortably. To be honest we could easily reduce the load by around 80% on our current server simply by adding a proper cache to a few legacy (before my time) applications. However at this point it all becomes political (which also reduces my hopes of us using Postgres).Where are the slides?!