News for hackers. Stuff that matters.
Re:Why Startups Die by derwiki on Wednesday October 03, 2012 @12:20PM (#4608175)
> Pre-Emptive Scaling

Man, this one a thousand times over. I shudder when I think of all the work that could have been saved if these premature optimizations had never been made (I wrote a blog post on 500m row tables in MySQL about this). A little dose of YAGNI would go a long way to many startups. Also, re: Twitter and Tumblr's scaling problems: they didn't fail because of sporadic downtime. They fixed the actual problems/bottlenecks when they encountered them, and went on.

Re:Why Startups Die by ftwinnovations on Wednesday October 03, 2012 @08:34PM (#4610358)
Agree 10,000x over your 1000x. On the occasions where rapid unexpected (and awesome!) traffic or growth spikes happen you can almost always temporarily patch the situation by scaling vertically (bigger/faster disk/CPU) or quick and dirty horizontally (load balancing or replication) and then have some breathing room to work out the real meaningful or cost effective scaling solution.