Large Site Network Infrastructure (Hardware) Examples?
February 16, 2008 4:03 PM Subscribe
Network admin / geek question:
I'm in the process of building a small cluster of servers for my quickly growing wallpaper website, but I've never charted those waters before. I found a nice article on TorrentFreak about Mininova's server setup, and found it quite intriguing. My question is, are there any other big-ish sites like Mininova that are open about their hardware infrastructure? Googling around yielded few results. I'd be interested to see the setup for a site like Digg, MeFi, or Flickr. Thanks in advance.
Currently, I have a single Dual P3 load balancer, four DL360 (each 2x 2.8Ghz xeon) servers for apache, one DL560 (quad 2.8 xeon) server for MySQL and NFS, and a few misc. servers.
Do most large sites use NFS to share the website source? Any idea how much load NFS puts on a server? Also, any idea how much overhead a load balancer requires to route connections? Can a dual P3 1.6Ghz handle a 15mbit'ish site?
Any links or advice would be appreciated.
Thanks.
Currently, I have a single Dual P3 load balancer, four DL360 (each 2x 2.8Ghz xeon) servers for apache, one DL560 (quad 2.8 xeon) server for MySQL and NFS, and a few misc. servers.
Do most large sites use NFS to share the website source? Any idea how much load NFS puts on a server? Also, any idea how much overhead a load balancer requires to route connections? Can a dual P3 1.6Ghz handle a 15mbit'ish site?
Any links or advice would be appreciated.
Thanks.
Oh, and so far you're doing pretty much what I've been doing for my high-availability setups ... c'ept I usually integrate squid somehow and also usually use memcached to store sessions instead of relying on nfs not to screw up. (But I don't like nfs that much... our setups are stable right now, but around 2000 I had some serious problems with some NFS setups and it soured me on the technology.) What I generally have is a set of dynamic servers with php/apache, and a set of static asset servers running lighttpd and just pushing content.
And NFS, if it's placed on a dedicated box (We usually use solaris for this due to ZFS) can push a gig a sec easy as long as it's got a processor core and I/O capable of that.
posted by SpecialK at 4:16 PM on February 16, 2008
And NFS, if it's placed on a dedicated box (We usually use solaris for this due to ZFS) can push a gig a sec easy as long as it's got a processor core and I/O capable of that.
posted by SpecialK at 4:16 PM on February 16, 2008
This recent thread is very much worth reviewing. Key link: http://www.highscalability.com/.
posted by idb at 5:31 PM on February 16, 2008
posted by idb at 5:31 PM on February 16, 2008
Read plenty on Amazons setup. If you like free, you can L4 with ipfilter, and l4ip. :)
posted by lundman at 7:18 PM on February 16, 2008
posted by lundman at 7:18 PM on February 16, 2008
One of the Youtube people gave a talk about their infrastructure. Here's the video of the talk.
posted by zippy at 11:58 PM on February 16, 2008
posted by zippy at 11:58 PM on February 16, 2008
« Older What features should I add to my basic backup... | Where do I look to find information on rail cargo... Newer »
This thread is closed to new comments.
What you might look into after the load balancer is adding a caching layer that uses Squid in reverse-proxy mode to take the load off of the apache servers.
Wikipedia is very open about their structure -- since they're a nonprofit. ... that's the largest open site I know. Another good source is Livejournal's Brad Fitzpatrick and his presentations on how they scaled their organization up to zomgwtfbbq.
posted by SpecialK at 4:09 PM on February 16, 2008 [1 favorite]