Slideshare.net (beta)

 
Post: 
Myspace Hi5 Friendster Xanga LiveJournal Facebook Blogger Tagged Typepad Freewebs BlackPlanet gigya icons



All comments

Add a comment on Slide 1

If you have a SlideShare account, login to comment; else you can comment as a guest


Showing 1-50 of 25 (more)

Scaling Twitter - Railsconf 2007

From al3x, 1 year ago

A conversational presentation on Twitter's scaling success with Ra

8406 views  |  4 comments  |  24 favorites  |  326 downloads  |  8 embeds (Stats)
 

Tags

twitter railsconf railsconf07 scaling rails rubyonrails al3x anotherbritt alexpayne brittselvitelle

more

 
 
 
 

Privacy InfoNew!

This slideshow is Public

 
Embed in your blog
Embed (wordpress.com)
custom

Slideshow Statistics
Total Views: 8406
on Slideshare: 8392
from embeds: 14* * Views from embeds since 21 Aug, 07

Slideshow transcript

Slide 1: Scaling Twitter Alex & Britt (from Twitter)

Slide 2: Preconceived notions.

Slide 3: The story.

Slide 4: So, what happened?

Slide 5: What we did.

Slide 6: More boxes. Not so many then. But now... ‣ Mongrels across 19 cores. ‣ Message processing/delivery across 16 cores. ‣ Jabber across 2 cores. ‣ MySQL on one big 8 core box. ‣ 16GB+ for memcache across a bunch of ‣ machines. ‣ 32 cores total.

Slide 7: That’s a lot.

Slide 8: Why? 200 - 300 connections per second. ‣ Spiking at 800 connections per second. ‣ We’ve done 11,000 connections per second. ‣ MySQL has spiked to 2,400 queries per second. ‣ Alexa says we have a whole mess o’ traffic, and ‣ they’re not even seeing the API traffic.

Slide 9: memcache.

Slide 10: memcache.

Slide 11: memcache.

Slide 12: Cache the hell out of everything.

Slide 13: Denormalize.

Slide 14: find(:all, :conditions => \"substring( truncate(id,0),-2,1) = #{@fugly_dist_idx}\", :limit => 10)

Slide 15: Starling.

Slide 16: Community.

Slide 17: YOU CAN HAS QUESTIONS