NYCPHP Meetup

[Rob Marscher]

On Mar 7, 2011, at 2:41 PM, Paul A Houle wrote:
> It's not cpu cycles that matter here,  but I/O bandwidth,  seek time,  and RAM for cache on your  database server.
...
> it's remarkably easy to put enough data into a database that you'll have performance problems,  particularly if you're importing data from public sources or if you develop a successful community site.

Agreed... I/O bottlenecks are a pain.  But still, things are on the up.  RAM is cheap and I/O bandwidth and seek times seem to be getting major upgrades later this year with cheaper SSD drives and technology like Intel Thunderbolt.

The Facebook engineering team has had some nice posts recently that I think are relevant to this conversation.

Being a "full-stack programmer":
http://www.facebook.com/note.php?note_id=461505383919

Man... I can't find it for the life of me right now, but they had a post detailing how they generate unique ids for all content across their entire system.  If I remember right, it's just a super powerful mysql cluster with an auto-increment bigint column.  All of their different pieces make a REST request to the id system to get their unique id.  But I could be wrong on the exact details since I can't find the original post.