I had a solo SaaS I ran from '06 to '15, operating on a 17 server cluster at, of all places, the former Enron data center in Los Angeles. In addition to the "tr

I had a solo SaaS I ran from '06 to '15, operating on a 17 server cluster at, of all places, the former Enron data center in Los Angeles. In addition to the "traditional" 3 tiers of Dev, Staging, and Production, we (the startup was 2, me and another) had production setup with redundancy. If some hardware failed, other portions of the cluster would re-route and/or assume the failed hardware's duties. The only single point of failure we had was a Federal Reserve quality hardware firewall - that was the best investment I made, as it sustained massive DDOS attacks and more without breaking a sweat.